Preserving Digital Materials

Preserving Digital Materials
Ross Harvey
K. G. Saur
Ross Harvey
K G Saur Mnchen 2005
Bibliographic information published by Die Deutsche Bibliothek Die Deutsche Bibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data is available in the internet at http://dnb.ddb.de. U Printed on acid-free paper 2005 K . G . Saur Verlag GmbH, Mnchen All Rights Strictly Reserved No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without permission in writing from the publisher Typesetting by Florence Production Ltd., Stoodleigh, Devon, Great Britain. Printed and bound by Strauss GmbH, Mrlenbach, Germany. ISBN 3 - 598 -11686-1
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Contents
Preface Chapter 1 What is Preservation in the Digital Age? Changing Preservation Paradigms Introduction Changing paradigms The need for a new preservation paradigm Changing denitions Preservation denitions in the digital world What exactly are we trying to preserve? How long are we preserving them for? What strategies and actions do we need to apply when we preserve them? Conclusion Chapter 2 Why do we Preserve? Who Should do it? Introduction Why preserve digital materials? Professional imperatives Other stakeholders The extent of the preservation problem How much data have we lost? Current state of awareness of the digital preservation problem Conclusion Chapter 3 Why Theres a Problem: Digital Artifacts and Digital Objects Introduction Modes of digital death
xi 1 1 2 3 7 9 14 15 16 16 17 17 18 19 21 24 26 31 33 35 35 37
vi
Contents
Digital artifacts Magnetic media Optical disks A future for digital artifacts? Digital objects more than digital artifacts Loss of functionality of access devices Loss of manipulation and presentation capabilities Weak links in the documentation chain and loss of contextual information Conclusion Chapter 4 Selection for Preservation The Critical Decision Introduction Selection for preservation, cultural heritage, and professional practice Selection criteria traditionally used by libraries Appraisal criteria traditionally used by archives Why traditional selection criteria do not apply to digital materials IPR, context, stakeholders, and the continuum approach Intellectual property rights and legal deposit Legal deposit legislation Context and community Stakeholder input Value of the continuum approach Developing selection frameworks for preserving digital materials Towards a new decision-making framework How much to select? Conclusion Chapter 5 What Attributes of Digital Materials Do We Preserve? Introduction Digital materials, technology and data The importance of preserving context The OAIS reference model The role of metadata Persistent identiers Authenticity Signicant properties Research into authenticity Functional Requirements for Evidence in Recordkeeping Project (Pittsburgh) InterPARES 1 and 2 Trusted digital repositories Conclusion Chapter 6 Overview of Digital Preservation Strategies Introduction Historical overview Who is doing what?
39 41 43 44 46 46 48 49 51 53 53 54 57 59 61 62 63 63 64 65 66 67 69 72 73 75 75 77 80 81 83 87 88 91 95 95 96 97 98 99 99 102 103
Contents vii
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
The Australian experience Criteria for effective strategies and practices Typologies of principles, strategies and practices A typology of digital preservation? Conclusion Chapter 7 Preserve Technology Approaches: Tried and Tested Methods Introduction Non-solutions Do nothing Storage and handling practices Durable/persistent digital media Analogue backups Policy development Standards Digital archaeology Preserve technology approaches Technology preservation Technology watch Emulation The Universal Virtual Computer Conclusion Chapter 8 Preserve Objects Approaches: New Frontiers? Introduction Preserve Objects approaches Bit-stream copying, refreshing, and replication Bit-stream copying Refreshing Replication Standard data formats File format registries Standardizing le formats Restricting the range of le formats Developing archival le formats XML Migration Viewers and migration on request Encapsulation Digital mass storage systems Conclusion: combining principles, strategies and practices Chapter 9 Digital Preservation Initiatives and Collaborations Introduction Collaboration Typologies of digital preservation initiatives International initiatives and collaborations International services The Internet Archive
105 107 110 115 116 117 117 118 120 120 122 123 125 125 126 126 127 128 129 133 134 135 135 136 137 137 137 138 139 141 143 144 145 146 147 153 153 154 156 157 157 157 159 161 161 161
viii
Contents
The International Internet Preservation Consortium JSTOR DSpace LOCKSS OceanStore International alliances UNESCO RLG PADI OCLC CAMiLEON Regional initiatives and collaborations Regional services NEDLIB Regional alliances ERPANET Digital Recordkeeping Initiative National initiatives and collaborations National services KBs e-Depot National Archives Digital Curation Centre AHDS UK Web Archiving Consortium National alliances Digital Preservation Coalition NDIIPP Digital Preservation Testbed Sectoral initiatives and collaborations Sectoral services PARADISEC CEDARS Sectoral alliances JISC Conclusion Chapter 10 Challenges for the Future of Digital Preservation Introduction What have we learned so far? Four major challenges Challenge 1: managing digital preservation Challenge 2: funding digital preservation Challenge 3: peopling digital preservation Challenge 4: making digital preservation t Research and digital preservation Conclusion the future of digital preservation Appendix Case Studies Case Study 1: Digital Storage at the National Film and Sound Archive, Australia The standard
163 163 164 165 166 166 166 167 168 168 168 169 169 169 169 169 170 171 171 171 172 172 173 174 174 174 175 176 177 177 177 178 178 178 179 181 181 182 185 185 187 189 191 193 196 199 199 200
Contents
ix
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Functions and requirements The media Strategies Case Study 2: PANDORA (Preserving and Accessing Networked Documentary Resources of Australia) Case Study 3: Changing Preservation Activities at the National Archives of Australia Case Study 4: VERS (Victorian Electronic Records Strategy) Case Study 5: Migration at the Department of Family and Community Services, Australia Case Study 6: PADI (Preserving Access to Digital Information)
201 202 203 203 208 212 215 219
Bibliography Index
221 238
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Preface
Preserving Digital Materials investigates the current practice of librarians and recordkeeping professionals in digital preservation. A strong claim can be made that preservation of digital materials is the single most critical issue faced by members of the library and recordkeeping professions. Much information about digital preservation is available in print and on the web, and the quantity is increasing, but most practitioners do not have the time or the technical expertise to evaluate and synthesize it. Few single-volume guides to the complex issues surrounding digital preservation exist; most books about the preservation of library and archival material are largely concerned with the preservation of physical artifacts and collections and pay little attention to the serious issues raised by digital materials. Preserving Digital Materials lls this gap in the literature. It provides an introduction to the principles, strategies and practices applied by librarians and recordkeepers to the preservation of digital materials. It aims to improve digital preservation practice in libraries and recordkeeping environments. It does this by taking stock of what we know about the principles, strategies and practices that prevail and by describing the outcomes of recent and current research. The incorporation of practice from both the recordkeeping and the library communities, to which is added the experiences of other relevant communities such as the geoscience community, provides some new ways of viewing and thinking about aspects of digital preservation. Christine Borgman notes that digital preservation looms as one of the greatest challenges of information management technology and policy (Borgman, 2000, p.66). The problem has been more colourfully described as the dark archival underbelly of the wonderful world of digitized information and on line everything (US-InterPARES Project, 2002, p.1, quoting the Boston Globe, 10 February 2000). The National Library of Australias Digital Preservation Policy neatly sums up the challenges to ongoing accessibility of digital resources:
xii
Preface
The volume and growth in the amount of material to be maintained Widespread use of relatively unstable media Rapid changes in the availability of hardware, software and other technology required for access The diverse and frequently changing range of le formats and standards Uncertainty about the signicant properties that must be maintained for different digital resources The recurrent, critical and demanding nature of the treats to accessibility Uncertainty about strategies and techniques for addressing the threats The high costs of taking action Administrative complexities in ensuring timely and cost-effective action is taken over very long periods of time New models of ownership that impose property and other rights-based constraints The as-yet ambiguous intentions and capabilities of a range of agents with a potential impact on accessibility (National Library of Australia, 2002). Failure to address these challenges will result in the loss of signicant quantities of digital information. The paradigms that shape the environment in which we now function have changed. In comparison with traditional library and recordkeeping practice, signicantly different issues are being raised by the increasing reliance in todays society on information in digital form, both born digital and digitized from paper-based, lm-based, or other media. The benign neglect that may have sufced in the past to preserve information will no longer be enough: active intervention is required. The preservation of digital materials poses many challenges for which predigital paradigms offer little assistance. One challenge is that preserving digital materials requires constant maintenance, relying on complex hardware and software that are frequently upgraded or replaced. Another is the increasing quantity and complexity of digital materials, which tax current preservation strategies designed to archive relatively simple and self-contained collections of data and documents (Workshop on Research Challenges in Digital Archiving and Long-term Preservation, 2003, pp.viiiix). The range of stakeholders who have an interest in maintaining digital materials for use into the future is wide. A 2003 report begins with the words The need for digital preservation touches all our lives, whether we work in commercial or public sector institutions, engage in e-commerce, participate in e-government, or use a digital camera. In all these instances we use, trust and create e-content, and expect that this content will remain accessible to allow us to validate claims, trace what we have done, or pass a record to future generations (NSF-DELOS Working Group on Digital Archiving and Preservation, 2003, p.i). We cannot expect a technological quick x: the challenges of maintaining digital archives over long periods of time are as much social and institutional as technological (Workshop on Research Challenges in Digital Archiving and Long-term Preservation, 2003, pp.viii-ix). The authors of Preservation Management of Digital Materials: A Handbook see the issues as technological, organizational and legal (Jones & Beagrie, 2001, section 2.2). These and many other factors
Preface
xiii
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
combine to make the task difcult, with at present only a few clear pointers to the way ahead. Both the library community and the recordkeeping community (archivists and records managers) have energetically sought solutions to the challenges of digital preservation, but to date there has been relatively little sharing of outcomes. For example, much of the archives professions thinking about dening the authenticity of records in digital form is little known to the library community, as is the continuum approach to archives and records (a concept closely associated with Australian recordkeeping practice), which allows an integrated approach to managing electronic records and has proved to be a powerful tool for dealing with electronic records. Developments such as these have considerable potential to assist practice in other information and heritage communities, who need to learn more about them. This book goes some way towards addressing this need by providing examples from recordkeeping practice. Although much high-quality information is available to information professionals concerned with preserving materials in digital form, most notably on the web, its sheer volume causes problems for busy information professionals who wish to understand the issues and learn about strategies and practices for digital preservation. Preserving Digital Materials is written for librarians and recordkeepers. Its synthesis of current information, research, and perspectives about digital preservation from a wide range of sources across many areas of practice makes it of interest to a wide range of information professionals from preservation administrators and managers, who want a professional reference text, to thinking practitioners, who wish to reect on the issues that digital preservation raises in their professional practice. It will also be of interest to students in these elds. Preserving Digital Materials is not a how-to-do-it manual, so is of little assistance for those who want to learn how to carry out technical procedures. It is not primarily concerned with digitization and makes no distinction between information that is born digital and information that is digitized from physical media. This book addresses four key questions which give the text its four-part structure: 1. 2. 3. 4. Why do we preserve digital materials? What digital materials do we preserve? How do we preserve digital materials? How do we manage digital preservation?
Chapters 1 to 3 address the rst question: why do we preserve digital materials? They examine key denitions and their relationship to ways of thinking about digital preservation, note some of the reasons why preservation is a strong professional imperative for librarians and recordkeepers, indicate the extent of the preservation problem for digital materials, and look at the reasons why a digital preservation problem exists. The question what digital materials do we preserve? is noted in chapters 4 and 5. Chapter 4 examines the issues of selection of digital materials for preservation, and chapter 5 notes the questions relating to what attributes of digital materials we need to preserve. The question how do we preserve digital materials? is covered in chapter 6, 7 and 8. An overview of digital preservation strategies is provided in chapter 6, and chapters 7 and 8 describe specic strategies. Aspects of the question how do
xiv Preface
we manage digital preservation? are noted in chapters 9 and 10. Chapter 9 describes the major digital preservation initiatives and collaborations. Chapter 10 examines some of the issues that digital preservation faces in the future. The Appendix, comprising six case studies, provides examples of Australian digital preservation practice that illustrate a range of activities in organizations of varying sizes and with varying resources. The reader should be aware that this book presents a Western view of preservation, a view not necessarily embraced by people of all cultures. The reader should also note the words in Article 9 of the UNESCO Charter on the Preservation of the Digital Heritage: The digital heritage is inherently unlimited by time, geography, culture or format. It is culture-specic, but potentially accessible to every person in the world. Minorities may speak to majorities, the individual to a global audience. The digital heritage of all regions, countries and communities should be preserved and made accessible, creating over time a balanced and equitable representation of all peoples, nations, cultures and languages (UNESCO, 2004). Preserving Digital Materials uses many Australian examples. Australia has had and continues to have a major inuence on the rest of the world in digital preservation. Australian practice in digital preservation from the library, recordkeeping, audiovisual archiving, data archiving and geoscience sectors is often at the forefront of world best practice. The contention that many Australian digital preservation initiatives are considered as international best practice can be readily demonstrated. One example is Beagries 2003 report National Digital Preservation Initiatives (Beagrie, 2003), written to advise the Library of Congress as it develops its NDIIPP (National Digital Information Infrastructure and Preservation Program). Beagrie, at that time director of the United Kingdoms JISC Digital Preservation Focus, considered initiatives in Australia, France, the Netherlands, and the United Kingdom countries selected because their digital preservation activities were considered to be current best practice. Of Australia Beagrie notes: For a country with a relatively small population, Australia has a relatively large number of leading-edge online projects across all sectors. Archiving these online materials has become a signicant area of effort for Australias memory institutions, and both the NLA [National Library of Australia] and the national archive activities and guidelines are frequently cited internationally as exemplars in this area . . . There is an active electronic records management/archive sector in Australia. Work at Monash University, the PRO [Public Record Ofce] of Victoria, and the National Archive of Australia has earned an international reputation (Beagrie, 2003, p.14). The National Library of Australia provides a good example of Australian activity in digital preservation. A senior member of its executive could claim, after a study tour of national libraries in 2002, that the National Library of Australia is still in the forefront of digital archiving and preservation activities among national libraries (Gatenby, 2002b). In the 1980s, in common with many other
Preface
xv
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
research libraries, it began acquiring digital information resources whose value and signicance required their long-term retention. However, unlike most other research libraries, the National Library of Australia has been active in digital preservation since at least 1994, practically at the dawn of history in digital preservations timeline. It founded one of the worlds rst library digital preservation sections in 1995 and an archive of online publications, PANDORA, in 1996, again one of the rst in the world. Furthermore, it has actively shared its experience and expertise with the rest of the library community, not just within Australia but internationally. In fact, much of the National Librarys experience is available to anyone who cares to locate it on the Librarys web site. As one example, the Librarys Digital Archiving and Preservation at the National Library web page (www.nla.gov.au/initiatives/digarch.html) provides an outline of its digital preservation activities and links to its numerous policies, staff conference presentations, and relevant web sites. The National Librarys PANDORA Project has been acknowledged as deserving particular mention and commendation for the open and collegial atmosphere in which it is being undertaken, an approach which encourages shared learning and dialogue (Brindley, 2000, p.130), and it is closely observed to see how feasible the approach is (Smith, 2003, p.22). The National Librarys activities are applied to a wide range of concerns: the preservation of Australian online publications, archiving web sites, legal deposit, a web subject gateway for preservation resources, its Digital Services Project, persistent identiers, metadata, research into migration, a digital archiving system, authenticity, intellectual property rights management, digital audio-video archiving, and guides and codes of practice for creators and publishers of digital material. The National Librarys experience and expertise are frequently sought out, one recent example being authorship by its staff of the UNESCO Guidelines for the Preservation of Digital Heritage (UNESCO, 2003). Not only is the Librarys expertise sought, but there has been export of talent to the United Kingdom: Deborah Woodyard was, until recently, at the British Library and Maggie Jones is now at the Digital Preservation Coalition. It is not just the National Library of Australias activities that cause so much Australian digital preservation practice to be described as international best practice. From the National Archives of Australia have come theoretical and conceptual developments that have assisted thinking in digital preservation. One example is the National Archives of Australias concept of non-custodial or distributed-custody archives. The signicance of the distributed custody model lay in the heightening of awareness that the responsibility for preservation is not solely that of the archives, but with digital records must be considered at the point when the record is created. The National Archives of Australia has since reverted to a custodial model in which digital records are moved to the National Archives and converted to XML (Heslop, Davis & Wilson, 2002). Another important conceptual development originating in the Australian recordkeeping community is the continuum model developed by Sue McKemmish and Frank Upward at Monash University. This model has allowed changes in thinking about custody of records who owns the records, and who should therefore take the responsibility for their preservation (Upward, 2005). The Recordkeeping Metadata Project at Monash University has also been inuential. Another example from the Australian recordkeeping community is VERS (the Victorian Electronic Records Strategy), which aims to reliably and authentically
xvi Preface
archive electronic records created or managed by the Victorian government (www.prov.vic.gov.au/vers). Signicant expertise also resides in institutions outside the cultural heritage institutions, such as CSIRO (the Commonwealth Scientic and Industrial Research Organisation), the Australian nuclear industry, and the defence sector. Geoscience Australia, for example, manages a digital archive of over 700 terabytes and has carried out large-scale data recovery from tapes. The Australian Defence Force Academys Authenticated Australian Editions project is addressing issues, such as authenticity of digital objects, that cultural institutions need to know more about (idun.itsc.adfa.edu.au/ASEC/ aueledns.html). As noted above, there is a considerable amount of high-quality information available about preserving materials in digital form, and much of it is available on the web. The accessibility that this provides is countered by the impermanence of much web material, as noted in several chapters in this book. All URLs in this book were correct at the time of writing. Producing a book of this nature involves the input of many, and I have beneted from discussions with many colleagues over the last three years. In particular, I wish to acknowledge the following individuals for their ideas and support: Tony Dean for suggesting the example of Piltdown Man; Liz Reuben, Matthew Davies, Stephen Ellis and Rachel Salmond for case studies; Alan Howell, of the State Library of Victoria, and staff of the National Library of Australia, in particular Pam Gatenby, Colin Webb, Kevin Bradley, and Margaret Phillips, for their assistance with clarifying concepts. Some of the material in this book is based on interviews with Australian digital preservation experts, whose assistance and encouragement was invaluable. Alan Howell, Heather Brown and Peter Jenkins provided examples used in this book, and their assistance and the permission of the State Library of South Australia is gratefully acknowledged. Thanks are due to Ken Thibodeau, CLIR, ERPANET, and UNESCO for permission to use their material. I am grateful to my employer, Charles Sturt University, who supported me by providing study leave in 2003. I was fortunate to be based at the National Library of Australia as a National Library Fellow from March to June 2003, and I would like to thank its Director General, Jan Fullerton, and other staff of the National Library for this generous support. Finally, I acknowledge the unfailing support of Rachel Salmond in this and others of my endeavours.
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Chapter 1
What is Preservation in the Digital Age? Changing Preservation Paradigms
Introduction
Digital images are produced without the intermediaries of lm, paper or chemicals and as such never acquire the burden of being originals because they do not pass through a material phase (Sassoon, 1998, p.5, quoting Bruce, 1994) To preserve, as the dictionary reminds us is to keep safe . . . to maintain unchanged . . . to keep or maintain intact. But the rapid obsolescence of information technology entails the probability that any digital object maintained unchanged for any length of time will become inaccessible (Thibodeau, 1999) Any discussion about the preservation of digital information must begin with the consideration of two interlinked areas: changing preservation paradigms, and denitions of terms. Without a clear concept of what we are discussing the potential for confusion is too great. In library and recordkeeping practice we are moving from collection-based models, whose principles and practices have been developed over many centuries, to models where collections are not of paramount importance and where what matters is the extent of access provided to information resources, whether they are managed locally or remotely. As these models are not yet commonly agreed upon or applied, the situation is one of considerable ux. Archivists, for example, have considered, debated and sometimes applied the concept of non-custodial archives, where there is no central collection, to accommodate the massive increase in growth of digital records. Librarians manage hybrid libraries, consisting of both physical collections and distributed digital information resources, and virtual libraries. Other stakeholders with a keen interest in digital preservation manage digital information in specic subject areas, such as geospatial data or social science data. In the past this material, where it existed, was maintained as collections of paper and other physical objects.
2 What is Preservation in the Digital Age? Changing Preservation Paradigms
These changing models of library and recordkeeping practice are generating a need for new denitions. The old terms do not always convey useful meanings in the digital environment and can, in fact, be misleading, and, on occasion, harmful. In library and recordkeeping practice we are changing from a preservation paradigm where primary emphasis is placed on preserving the physical object (the artifact as carrier of the information we wish to retain), to one where there is no physical carrier to preserve. What, for example, does the term preservation mean in the digital environment? How has its meaning changed? What are the implications of these changes? The phrase benign neglect provides an example of a concept that is helpful in the pre-digital preservation paradigm, but is harmful in the new. It refers to the concept that many information carriers made of organic materials (most notably paper-based artifacts) will not deteriorate rapidly if they are left undisturbed. For digital materials this concept is positively harmful. One thing we do know about information in digital form is that actions must be applied almost from the moment it is created, if it is to survive. Nor are the earlier paradigm denitions sufcient to accommodate new forms, such as works of art that have digital technologies as part of them, or time-dened creative enterprises such as performance art. This chapter examines the effect of digital information on traditional librarianship and recordkeeping paradigms and notes the need for a new preservation paradigm in an environment that is dynamic, highly changeable, and has many stakeholders, often with competing interests. It considers the differences between born digital and digitized information, and denes key terms.
Changing paradigms
It is now commonplace to hear or read that we live in an information society, one of whose primary characteristics is the widespread and increasing use of networked computing, which in turn relies on digital data. This is revolutionizing the way in which large parts of the worlds population live, work and play, and how libraries, archives, museums, and other institutions concerned with documentary heritage function and are managed. New expectations of these institutions are evolving. The signicance of this change is readily illustrated by just one example. The internet is rapidly becoming the rst choice for people who are searching for information on a subject, and a new verb, to google (derived directly from Google, the name of a widely used internet search engine) has entered our vocabulary, as illustrated in this quote: students prefer Googling to using search tools in the librarys online catalogs (George, 2003, p.6). The sheer size and rate of growth of the internet mean that, currently, no systems have been developed to provide comprehensive access to it. The systems that do exist are embryonic and experimental, and the quality of the information available on the web is variable. Attempts to estimate the rate of the internets growth have included counting the number of domain names over several years. There has been a dramatic increase since 1988 (see Figure 1.1). These major changes it is not too extreme to call it a revolution raise the question of how we keep the digital data we decide is worth keeping. The everincreasing quantities being produced do not assist us in nding an answer: Large organizations routinely add terabytes of storage capacity, and more and
The need for a new preservation paradigm 3
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Internet Domain Survey Host Count 350,000,000 300,000,000 250,000,000 200,000,000 150,000,000 100,000,000 50,000,000 0 Jan-94 Jan-95 Jan-96 Jan-97 Jan-98 Jan-99 Jan-00 Jan-01 Jan-02 Jan-03 Jan-04 Jan-05
Source: Internet Software Consortium (www.isc.org)
Figure 1.1 Growth Rate of the Internet

(From www.isc.org/ds/, viewed 24 March 2005)
more individuals can afford laptop and desktop computers with tens of gigabytes of storage (Workshop on Research Challenges in Digital Archiving and Long-term Preservation, 2003, p.vii). Nor does the rapidity with which changes in computer and information technology occur. The challenges are new and complex for nearly all aspects of librarianship and recordkeeping, including preservation. There has also been a change in the ways in which information is produced and becomes available to its community of users. The internet is only one of these ways. In the pre-digital (print) environment the processes of creation, reproduction and distribution were separate and different; now, technology tends to erase distinctions between the separate processes of creation, reproduction and distribution that characterize the classic industrial model of print commodities (Nunberg, 1995, p.21). This has signicant implications for preservation, especially in terms of who takes responsibility for it and at what stage preservation actions are rst applied. For instance, in the industrial-mode print world, securing (acquisition) of the artifact the book so that it could be preserved occurred by means such as legal deposit legislation, requiring publishers to provide copies to libraries for preservation and other purposes. If the creator is now also the publisher and distributor, as is often the case in the digital world, who has the responsibility of securing (acquiring the information)? These points are noted in more detail later in this book.
The need for a new preservation paradigm

The digital revolution is changing the professional practice of librarians, recordkeepers, and all other information professionals. The paradigm shift in how
libraries, archives and information agencies conduct their activities can be crudely and naively described as the change from acquiring, storing and providing access to information resources in physical forms, to acquiring, storing and providing access to digital information resources. Preservation activities are also part of this new paradigm. The principles that lie behind preservation practices, and the techniques themselves, are not exempt. (Webb (2000a, 2000b) provides examples of the effects of these changes on preservation practice in Australia.) The pre-digital preservation paradigm is based on principles such as the following (based on Cloonan (1993, p.596) and Harvey (1993, pp.14,140): When materials are treated, the treatments should, when possible, be reversible Whenever possible or appropriate, the originals should be preserved; only materials that are untreatable should be reformatted Library materials should be preserved for as long as possible Efforts should be put into preventive conservation, and aimed at providing appropriate storage and handling of artifacts Benign neglect may be the best treatment. The denitions associated with the old preservation paradigm are rmly rooted in the conservation of artifacts the physical objects that carry the information content. In fact, the term materials conservation is sometimes used. For example, in the museum environment denitions are phrased in words such as these: Materials Conservation includes all actions taken to safeguard objects and the information contained in them, and Preservation aims to stabilize an object in its current condition with the least possible intervention (AMOL, 2003). The denitions provided in the IFLA Principles for the Care and Handling of Library Materials (Adcock, 1998), widely adopted in the library and recordkeeping contexts, articulate principles rmly based on maintenance of the physical artifact. The denition of Conservation notes that its aims are to slow deterioration and prolong the life of an object, and that of Archival quality emphasizes the longevity and stability of materials in the words a material, product, or process is durable and/or chemically stable, that it has a long life, and can therefore be used for preservation purposes (Adcock, 1998, p.4). Medium/media is dened as the material on which information is recorded. Sometimes also refers to the actual material used to record the image (Adcock, 1998, p.5). The point here is that the old preservation paradigm considers information content and the carrier as one and the same, although this changed in the declining years of the old paradigm as large-scale copying programmes, especially microlming programmes, were implemented. To these denitions we should add Restoration (from the 1986 version of the IFLA Principles): Denotes those techniques and judgements used by technical staff engaged in the making good of library and archive materials damaged by time, use and other factors (Dureau and Clements, 1986, p.2). That pre-digital paradigm thinking does not transfer well to the digital environment is easily illustrated, as these two articles from one issue of an Australian librarianship journal indicate. One article describes the Parish Map Project, which is providing access, in full colour, to an important historical archive in New South Wales (Read, 1999, p.45). It is apparent from this article that preservation in this project is viewed as retaining the original paper artifacts and
The need for a new preservation paradigm 5
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
restricting their use, while providing digital copies for access. No mention is made of how the longevity and integrity of the digital copies might be ensured. The second article notes the First Families 2001 programme which aims to collect and preserve the life stories of Australias diverse cultural heritage in an online database (Van de Velde, 1999, p.5). Preservation is mentioned only three times: in the title, the abstract and on the rst page. The reader could easily deduce from the report on this project that collecting the material is the same as preserving it. Cox indicates how a different way of thinking is needed in the recordkeeping community and suggests four elements of a new preservation paradigm for electronic records. He notes, for instance, that recordkeepers have long seen centralization or custody of records as crucial to their work, but this is not feasible for electronic records: transferring electronic records to the custody of archives may undermine their very long-term use (Cox, 2001, p.95). As Marcum notes in her preface to a 2002 survey about the state of preservation programmes in American college and research libraries, the information landscape has changed, thanks to the digital revolution. Libraries are working to integrate access to print materials with access to digital materials. There is likewise a challenge to integrate the preservation of analog and digital materials. Preservation specialists have been trained to work with print-based materials, and they are justiably concerned about the increased complexity of the new preservation agenda (Kenney and Stam, 2002, p.v). What is the new preservation agenda? How has the preservation paradigm changed? Pre-digital preservation paradigm thinking does include some useful understanding of digital preservation. For example, it recognizes that copying (as in refreshing from tape to tape) is the basis of digital preservation; it does not, however, engender an understanding of the complexity of copying which is more than simply preserving a bit-stream, but must take account of a wide range of other attributes of the digital object that also need to be preserved. This recognition is encapsulated in British Standard BS4783 Part 2, dating from 1988, which focuses on procedures for refreshing digital data by copying it from magnetic tape to magnetic tape at regular intervals, and on the best storage and handling of these tapes. Hedstrom in 1998 provided early recognition of the need for new preservation paradigms and of the enormity of the challenges we face in digital preservation, noting that digital preservation adds a new set of challenges for libraries and archives to the existing task of preserving a legacy of materials in traditional formats (Hedstrom, 1998, p.192). However, old-paradigm preservation thinking has led to the characterization of some current thinking about digital preservation as a myopic focus on technical problems (such as preserving digital objects) and a concomitant neglect of the bigger picture (for example, public policy, among other issues) (Cloonan, 2001, p.232). Not only are there additional technical challenges, but there are also new challenges resulting from the quantities of digital information being produced. Hedstrom again: Our ability to create, amass, and share digital materials far exceeds our current capacity to preserve even that small amount with continuing value (Hedstrom, 1998, p.192). Specic examples from scientic areas, such as
determining long-term global climate change, and biomedicine, are provided in a 2003 report which contends that much more digital content is available and worth preserving (Workshop on Research Challenges in Digital Archiving and Long-term Preservation, 2003, p.2). New ways of thinking about preservation and new skills are needed. In essence the whole of this book is about new ways of thinking about preservation. At this point, however, it is useful to consider some of the key elements of new paradigm preservation thinking. The rst of these is the need to actively maintain digital information over time from the moment of its creation. Interruptions in the management of a digital collection will mean that there is no collection left to manage. Unlike most collections of physical objects, collections of digital materials require constant maintenance and elaborate lifesupport systems to remain viable (Workshop on Research Challenges in Digital Archiving and Long-term Preservation, 2003, p.7). In addition to these technical issues there are other issues social and institutional (political in the broadest sense) for even the most ideal technological solutions will require management and support from institutions that go through changes in direction, purpose, and funding (Workshop on Research Challenges in Digital Archiving and Long-term Preservation, 2003, p.7). Another key element is the scale and nature of the digital information we wish to maintain into the future, and the preservation challenges these pose. The complexities of the variety of digital materials are described in this way: Digital objects worthy of preservation include databases, documents, sound and video recordings, images, and dynamic multi-media productions. These entities are created on many different types of media and stored in a wide variety of formats. Despite a steady drop in storage costs, the recent inux of digital information and its growing complexity exceeds the archiving capacity of most organizations (Workshop on Research Challenges in Digital Archiving and Long-term Preservation, 2003, p.7). Arising from these key factors is the need for new kinds of skills. Current preservation skills and techniques are labour-intensive and, even where appropriate, do not scale up to the massive quantities of digital materials we are already encountering. The problem cannot simply be addressed by technological means: Digital collections require curation and processing to ensure their longevity, protect their integrity, and enhance their value for use in the future (Workshop on Research Challenges in Digital Archiving and Long-term Preservation, 2003, p.7). What kind of person will implement the new policies and develop the new procedures we require to maintain digital materials effectively into the future? New kinds of positions, requiring new skill sets are already being established in libraries and archives. Key selection criteria for a Digital Preservation Specialist at the State Library of Victoria, Melbourne, advertised in May 2003, included: Demonstrated knowledge, expertise and level of excellence in digital object management
Changing denitions
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Sound knowledge of, and practical competence with, computers and the nature of digital information, archiving of data, physical formats, computer systems and the Internet Familiarity with contemporary and superseded software and media and demonstrated understanding of the emerging eld of digital archiving, taking account of short and long term access and preservation needs of collection material. But more than new skills and new ways of thinking are required. We need to redene the eld of preservation and the terms we use to describe preservation activities.
Changing denitions
As already suggested, pre-digital preservation paradigm denitions do not convey useful meanings when they are applied to digital preservation. What are they? Currently conservation is the more specic term and is particularly used in relation to specic objects, whereas preservation is a broader concept covering conservation as well as actions relating to protection, maintenance, and restoration of library collections. The eminent British conservator, Christopher Clarkson, emphasizes this broader aspect when he states that preservation encompasses every facet of library life: it is, he says, preventive medicine . . . the concern of everyone who walks into, or works in, a library. For Clarkson conservation is the specialized process of making safe, or to a certain degree usable, fragile period objects and restoration expresses rather extensive rebuilding and replacement by modern materials within a period object, catering for a future of more robust use. He neatly distinguishes the three terms by relating them to the extent of operations applied to an item: restoration implies major alterations, conservation minimal and preservation none. (Harvey, 1993, pp.67) These denitions are based on the assumptions that deteriorated materials or artifacts can be made good by restoring them, and that we can slow down, perhaps even halt, the rate of deterioration by taking appropriate measures such as paying attention to careful handling and appropriate storage. The principles on which old-paradigm preservation practice was based, and in which these denitions are rooted, are almost entirely oriented towards artifacts the preservation of the media in which the information is stored. These principles were altered as large-scale mass preservation treatments and practices, such as mass deacidication or microlming, were implemented. For example, reformatting programmes (microlming and photocopying) moved the paradigm towards a recognition that information content could be preserved without the need also to preserve the original media. In a prescient conjecture about how these preservation principles will change in the future, Cloonan in 1993 asked Which principles will be practiced? . . . Will our very concept of permanence change in the next generation . . . ? The answer . . . is likely to be yes (Cloonan, 1993,
p.602). She suggests that our preservation concerns in the future will be about the preservation of knowledge, not the preservation of individual items, and that we must continue to save as much information as possible, regardless of the format or the means by which it is stored and disseminated (Cloonan, 1993, p.603). But these denitions simply do not work with digital materials. Why not? Cloonan suggests part of the answer the preservation of knowledge, not the preservation of individual items but there is much more. She suggests that, while it is easy to indicate what a library preservation programme consists of disaster recovery planning, collection development policies, environmental controls, integrated pest management, proper storage, physical treatment, reformatting, migration, staff and user education, and the like we are still not any closer to what preservation is really about, what its essence is. Is preservation merely a set of actions? Is it a way of seeing? Or a way of interpreting information? Is copying preservation? Is reformatting? (Cloonan, 2001, pp.232233). Others have described the changes from old to new preservation paradigms and the development of new denitions in terms of transformation. Conway suggests that consensus about a set of fundamental principles that should govern the management of available resources in a mature preservation program has been reached in the analogue world and are the same in the digital world. They in essence, dene the priorities for extending the useful life of information resources. These concepts are longevity, choice, quality, integrity, and accessibility (Conway, 2000, p.22). However, these concepts have been transformed to accommodate the preservation of digital information. Longevity has altered from a focus on extending the life of physical media to one on the life expectancy of the access system. Choice (selection of material to be preserved) is no longer a decision made later in the life cycle of an item, but for digital materials has become an ongoing process intimately connected to the active use of the digital les. Integrity, based on the authenticity, or truthfulness, of the information content of an item, has transformed from a primary emphasis on maintaining the physical medium, to developing procedures that allow us to be assured that no changes have been made: ultimately, the digital world fundamentally transforms traditional preservation principles from guaranteeing the physical integrity of the object to specifying the creation of the object whose intellectual integrity is its primary characteristic. In the digital world, access to the artifact is clearly no longer sufcient; what is required, suggests Conway, is access to a high quality, high value, well-protected, and fully integrated digital product (Conway, 2000, p.27). The changing paradigms and denitions also entail major shifts in the conceptualization of preservation. Cloonan (2001) refers to the paradox of preservation: it is impossible to keep things the same forever. To conserve, preserve, or restore is to alter (Cloonan, 2001, p.235). Preservation implies alteration. But, if we seek to keep a digital object unchanged, we are in effect requiring that the technology on which it was developed to operate is also maintained. Over time this raises increasing problems. We reduce access, therefore, to digital materials we are attempting to maintain by preserving them. An illustration of the problem is provided by Thibodeau, Moore and Baru:
Preservation denitions in the digital world 9
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
many users would not be pleased if, in order to access digital objects that had been preserved across the last 30 years, they had to learn to use the PL1 programming language or Model 204 database software (Thibodeau, Moore and Baru, 1999). One implication of this paradox is that we should accept some degree of change in digital materials as we preserve them over time. The real issue is: how much change should we accept? Transformations are also apparent in how information professionals dene preservation. Denitions by professional associations show some appreciation that something more than preserving the object is required (Cloonan, 2001, pp.235237). The Society of American Archivists, for example, maintains that preservation of digital information is not so much about protecting physical objects as about specifying the creation and maintenance of intangible electronic les whose intellectual integrity is their primary characteristic (Society of American Archivists, 1997). In a 2002 survey of records managers and archivists who deal with electronic records, Cloonan and Sanett posed the questions: What is the meaning of preservation? Does the meaning change when it is applied to electronic rather than paper-based records? (Cloonan and Sanett, 2002, p.73). They noted that: It is clear that professionals are revising their denitions of preservation from a once-and-forever approach for paper-based materials to an all-the-time approach for digital materials. Preservation must now accommodate both media and access systems . . . while we once tended to think about preserving materials for a particular period of time for example, permanent/durable paper was expected to last for ve hundred years we now think about retaining digital media for a period of continuing value (Cloonan and Sanett, 2002, p.93). Interviewees expressed dissatisfaction with the term digital preservation, suggesting that other terms such as long-term retention are more suitable (Cloonan and Sanett, 2002, p.74). This survey indicates that information professionals now think about maximizing useful life or preserving digital documents forever through emulation or forward migration, but without the emphasis on a specic number of years (Cloonan and Sanett, 2002, pp.8687). Denitions of preservation provided by survey respondents, Cloonan and Sanett conclude, overall demonstrate a shift taking place from dening preservation as a once-andforever approach for paper-based materials, to an all-the-time approach for digital materials (Cloonan and Sanett, 2002, p.85). This shift implies an acceptance that preservation may even begin before a record has been created.
Preservation denitions in the digital world

What, then, are viable denitions of preservation in a digital world? What concepts do these denitions need to accommodate? What difculties do we encounter in trying to develop denitions that will assist us in our attempts to preserve digital materials?
10
The rst point to note is the dynamic nature of the eld. We must be prepared to change our paradigms and the denitions that we develop from them. In a discussion of how archivists thinking can better inform digital preservation, Gilliland-Swetland comments that the paradigms of any of the information professions come up short when compared with the scope of the issues continuously emerging in the digital environment. An overarching dynamic paradigm that adopts, adapts, develops, and sheds principles and practices of the constituent information communities as necessary needs to be created (GillilandSwetland, 2000, p.v). Another important concept that our denitions need to accommodate is that of information being preserved, independent of the media in which it resides. This is now well accepted; indeed, old preservation paradigm practices were, in recent decades, well acquainted with it through microlming programs. It is no longer a fact that the original has more integrity and veracity than a copy (Cloonan, 2001, pp.236237); instead, in the digital world, we need to look further to dene what attributes of the digital object we wish to maintain over time. Denitions should also accommodate the social and organizational aspects of digital preservation, the public policy, economic, political, social or educational perspective in Cloonans words (Cloonan, 2001, p.238). Old paradigm denitions certainly recognized that there was more to preservation than the technical aspects the IFLA Principles suggest it also encompasses managerial and nancial considerations, . . . stafng levels [and] policies (Adcock, 1998, p.5) but denitions of digital preservation need to go considerably further. They must be extended because of yet another factor, the need to start preserving digital materials almost from the moment of their creation and, some suggest, even before they are created. Preservation, in the pre-digital paradigm, was usually applied retrospectively. Conservation procedures were applied to artifacts only after the artifact, or the information contained in or on it, had been deemed to be of signicance and therefore worth preserving for use in the future. For example, books printed before 1800 are typically considered to be signicant because they are the product of handcraft production techniques and, therefore, no two items are identical; and some artifacts are preserved because they attain iconic status an Australian example is the veneration of relics of the nineteenth-century folk hero, Ned Kelly. We could also rely on benign neglect, where lack of action did not usually harm the item (assuming certain factors were in play, such as low use) and did not signicantly affect the likelihood of its survival. This concept no longer works for digital materials. The 3.5-inch diskette, very common until recently, provides a good example. Information on it is likely to become unreadable for many reasons: the diskette may be stored in conditions too humid or too hot, the drives to read it may be superseded by newer technology and no longer be available, the driver software for that drive may no longer be found. After a period of time, the diskette becomes unusable. Active preservation needs to start close to the time of creation if there is to be any certainty that the digital information will be accessible in the future. To this list we can add other concepts. One is that for digital materials their preservation must be an integral element of the initial design of systems and
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
projects (Ross, 2000, p.13). However, this is not usually the case. Another is that digital materials exist in a bewilderingly large number of formats there is still little standardization. And yet another, this one possibly the most signicant to accommodate, is that the preservation of digital materials is much more than the preservation of information content or physical carrier alone: it is about preserving the intellectual integrity of information objects, including capturing information about the various contexts within which information is created, organized, and used; organic relationships with other information objects; and characteristics that provide meaning and evidential value (Gilliland-Swetland, 2000, p.29). Preserving the original bit-stream is only one part of the problem: equally important is the requirement to preserve the means of interpreting, reading and utilizing the bit stream (Deegan and Tanner, 2002). The difculties of denition are not helped by disciplinary differences: there are, for instance, differences in the way terms are used by archivists and librarians, although they are drawing closer in the digital environment, where the integrity and authenticity of digital objects is of mutual concern to both professions (Cloonan and Sanett, 2000, p.210). (Note that integrity and authenticity are terms deriving from the archival profession and were, until recently, not usually associated with the work of librarians.) However, these differences pale in comparison with the signicantly different denitions used by the IT industry. How IT professionals think about the long-term storage of digital data is a question that assumes some importance for digital preservation, because of the heavy reliance that information professionals place on their skills and services. There are abundant signs that the mindset of IT professionals is signicantly different when they think about preservation. Denitions of archive, archiving and archival storage give us some indication of their concerns. A selection of IT dictionaries from the shelves of the National Library of Australias Reference Collection in April 2003 and an internet search indicated that these terms were used in two ways: 1. The process of moving data to a different kind of storage medium: for example, archive The process of moving data stored on an online, direct access device to an ofine storage medium because its frequency of use and timeliness permit a delayed access to the data (Sochat and Williams, 1992, p.16) 2. The process of backing up data for long-term storage: for example, archive v To backup or make copies for long-term storage (Wyatt, 1990, p.27). Few of these denitions display any interest or concern with the reasons why long-term storage might be required, although one notable exception was located: archiving Long term storage of information on electronic media. Information is archived for legal, security or historical reasons, rather than for regular processing or retrieval (Gunton, 1993, p.11). Perhaps the mindset of IT professionals is better indicated by this excerpt: You detect data thats not needed online and move it an off-shore store. When someone wants to use it, go nd the off-line media and restore the data (Faulds and Challinor, 1998, p.280). There is no indication in these denitions of the period of time that longterm refers to, yet this is a crucial point for those who are concerned with
12
preservation. My employer, an Australian university, in a discussion of longterm storage facilities, suggests access for a number of years, or a considerable time (years) (Message on my.csu, Subject line: Information Update DIT Archiving Project, 11 May 2002). While it is perhaps not especially helpful to dene long-term in terms of a specic number of months or years, some awareness of the problems is required, such as that evident in the NSF-DELOS Working Groups report Invest to Save (NSF-DELOS Working Group on Digital Archiving and Preservation, 2003, p.2) where the OAIS (Open Archival Information System) denition of long-term is noted: A period of time long enough for there to be concern about the impacts of changing technologies, including support for new media and data formats, and of a changing user community, on the information being held in a repository. What, then, are the denitions currently being used? Denitions have, of course, changed as we come to know more about how to preserve digital objects. The example of time how long we want to preserve material for illustrates this point. Long-term, initially incorporating elements of old-paradigm thinking of indenitely, or as long as possible as in digital preservation means retaining digital image collections in a usable and interpretable form for the long term (Kenney and Rieger, 2000, p.135) is now more commonly dened in the preservation community in terms derived from the archival community, as the period during which the information remains of continuing value. Such a denition is more helpful than the commonly encountered phrase over time as in ensuring the integrity of information over time (Gilliland-Swetland, 2000, p.22), and digital preservation [is] the processes and activities which stabilize and protect reformatted and born digital authentic electronic materials in forms which are retrievable, readable, and usable over time (Cloonan and Sanett, 2002, p.95). These issues and concepts are covered in two of the most comprehensive of the recent publications on digital preservation, the inuential Preservation Management of Digital Materials: A Handbook (Jones and Beagrie, 2001) and the UNESCO Guidelines for the Preservation of Digital Heritage (UNESCO, 2003). Selected denitions from these two sources are presented in Figure 1.2. A new term is gaining currency with the establishment of the Digital Curation Centre in 2004, although it is too early to know whether it will become the standard term. Curation is dened as the actions needed to maintain digital research data and other digital materials over their entire life-cycle and over time for current and future generations of users. Implicit in this denition are the processes of digital archiving and preservation but it also includes all the processes needed for good data creation and management, and the capacity to add value to data to generate new sources of information and knowledge . . . ; it is the key to reproducibility and re-use . . . Digital curation . . . is about maintaining and adding value to, a trusted body of digital information for current and future (Digital Curation Centre, 2004). These denitions assist us by providing useful starting points for an extended discussion of digital preservation. In particular, they address some signicant questions that we need guidance on:
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Guidelines for the Preservation of Digital Heritage (UNESCO, 2003) Accessibility The ability to access the essential, authentic meaning or purpose of a digital object (p.161). Digital materials cannot be said to be preserved if access is lost. The purpose of preservation is to maintain the ability to present the essential elements of authentic digital materials (p.22) Authenticity The digital material is what Authenticity Quality of genuineness it purports to be. In the case of electronic and trustworthiness of some digital records, it refers to the trustworthiness of materials, as being what they purport to the electronic record as a record. In the be, either as an original object or as a case of born digital and digitized reliable copy derived by fully documented materials, it refers to the fact that whatever processes from an original (p.161) is being cited is the same as it was when it was rst created unless the accompanying metadata indicates any changes. Condence in the authenticity of digital materials over time is particularly crucial owing to the ease with which alterations can be made (p.9) Digital heritage Those digital materials that are valued sufciently to be retained for future access and use (p.161) Digital Materials A broad term Digital materials is generally used here encompassing digital surrogates created as as a preferred term covering items of a result of converting analogue materials digital heritage at a general level. In some to digital form (digitization), and born places, digital object or digital resource digital for which there has never been have also been used. These terms have and is never intended to be an analogue been used interchangeably and generically equivalent, and digital records (p.1) (p.21) Digital Preservation Refers to the series Digital preservation The processes of of managed activities necessary to ensure maintaining accessibility of digital objects continued access to digital materials for over time (p.161) . . . is used to describe as long as necessary. Digital preservation the processes involved in maintaining is dened very broadly for the purposes information and other kinds of heritage of this study and refers to all of the actions that exist in a digital form. In these required to maintain access to digital Guidelines, it does not refer to the use of materials beyond the limits of media digital imaging or capture techniques to failure or technological change . . . (p.10) make copies of non-digital items, even if that is done for preservation purposes . . . (p.21) Information Packages . . . Preservation depends on maintaining digital objects and any information and tools that would be needed in order to access and understand them. Together, these can be considered to form an information package that must be managed either as a single object or as a virtual package (with the object and associated information tools linked but stored separately) (p.40) Preservation program The set of arrangements, and those responsible for them, that are put in place to manage digital materials for ongoing accessibility (p.162). . . . is used to refer to any set of coherent arrangements aimed at preserving digital objects. (p.21) Figure 1.2 Selected Denitions (From Jones and Beagrie, 2001; UNESCO, 2003)
Preservation Management of Digital Materials (Jones and Beagrie, 2001) Access . . . continued, ongoing usability of a digital resource, retaining all qualities of authenticity, accuracy and functionality deemed to be essential for the purposes the digital material was created and/or acquired for (p.9)
14
What exactly are we trying to preserve? How long are we preserving them for? What strategies and actions do we need to apply when we preserve them?
What exactly are we trying to preserve?

One of UNESCOs thematic areas is culture, and, within that, heritage. Therefore the UNESCO Guidelines are primarily concerned with digital heritage, those digital materials that are valued sufciently to be retained for future access and use. Although this statement is general and does not assist us to decide precisely what it is we want to preserve, it does introduce the essential concept of selection, of deciding value in this case, that digital material on which high value is placed. This is the subject of Chapter 4. More specic in the UNESCO Guidelines and in Preservation Management of Digital Materials: A Handbook are the denitions of digital materials the specic digital items, objects or resources that we are concerned with. Both publications use digital materials, suggesting a high level of consensus; this term will, therefore, be adopted in this book, despite the suggestion of pre-digital paradigm thinking in the physicality of the word materials, as things with a physical presence. Both sets of denitions categorically state that they are not concerned with the use of digitizing of analogue materials as a preservation technique. Jones and Beagries handbook specically excludes the potential use of digital technology to preserve the original artifacts through digitisation, and the UNESCO Guidelines are equally adamant, stating that digital preservation does not refer to the use of digital imaging or capture techniques to make copies of non-digital items, even if that is done for preservation purposes. Why such statements are thought to be worth making so rmly is explained by a strongly-held perception prevalent in the information professions that digitizing of analogue materials, often photographs or paper-based material, is sufcient for preservation purposes. Such attitudes have been noted earlier in this chapter: the Parish Map Project (Read, 1999) created digital content from original paper artifacts, and descriptions of such projects are very commonly found in the literature. Not all information professionals seem to be aware of this distinction and of its consequences. One exception is Sassoon, who fervently argues that the lack of understanding about just what is happening when photographs are digitized, about from where photographic meaning emanates (Sassoon, 1998, p.5), results in the loss of crucial information and raises serious ethical issues about the destruction of information as a consequence of digitization. In terms of how they are preserved, though, the denitions in these two sources make no distinction between born digital materials and digital materials created by digitizing analogue materials. This is acknowledged in Jones and Beagries denition of digital materials, which covers both digital surrogates created as a result of converting analogue materials to digital form (digitisation), and born digital for which there has never been and is never intended to be an analogue equivalent, and digital records (Jones and Beagrie, 2001, p.1). But it is not simply the bit-stream that we are seeking to preserve, as these denitions make clear. In order to ensure access in the future to digital materials, we also need to take account of other attributes of the digital materials. The UNESCO Guidelines indicate this in the denition of information packages.
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Guidelines for the Preservation of Digital Heritage (UNESCO, 2003) Conceptual objects Digital objects as humans interact with them in a humanunderstandable form (p.161) Essential elements The elements, characteristics and attributes of a given digital object that must be preserved in order to re-present its essential meaning or purpose. Also called signicant properties by some researchers (pp.161162) Logical objects Digital objects as computer encoding, underlying conceptual objects (p.162) Physical objects Digital objects as physical phenomena that record the logical encoding, such as polarity states in magnetic media, or reectivity states in optical media (p.162) Figure 1.3 Selected Denitions (From UNESCO, 2003)
In addition to the bit-stream, which is typically not understandable or re-presentable by itself, any information and tools that would be needed in order to access and understand the digital materials must also be preserved (UNESCO, 2003, p.40). The denitions are also very specic about the need to maintain other attributes of digital materials. To ensure that digital materials remain usable in the future, access to them is required and not simply access, but access to all qualities of authenticity, accuracy and functionality (Jones and Beagrie, 2001, p.9). This, in turn, requires denitions of authenticity, well expressed by the UNESCO Guidelines as the quality of genuineness and trustworthiness of some digital materials, as being what they purport to be, either as an original object or as a reliable copy derived by fully documented processes from an original (UNESCO, 2003, p.161). (Note the emphasis on the signicance of full documentation to ensure authenticity: this has important implications for digital preservation, noted further in Chapter 5.) Four further denitions in the UNESCO Guidelines clarify and emphasize these requirements (see Figure 1.3). Digital materials can be considered as physical objects, logical objects, or conceptual objects. The physical object is the artifact (for example, the diskette, the CD, or the magnetic tape whose physical characteristics store in or on it the bitstream that is, the logical object). These are given sense when they are used by humans and are labeled conceptual objects: what we deal with in the real world (Thibodeau, 2002, p.8). Essential elements (or signicant properties) of digital materials, when taken together, enable us to re-present the materials in the manner in which they were originally intended; that is, to preserve them. These concepts are noted further in the UNESCO Guidelines (see, for example, UNESCO, 2003, p.36) and elsewhere (Thibodeau, 2002, p.613).
How long are we preserving them for?

Although the denitions in the UNESCO Guidelines are not very helpful in providing guidance about the length of time we preserve digital materials, Jones and Beagrie assist us with their articulation of long-term, medium-term and short-term preservation. Long-term preservation aims to provide indenite access to digital materials, or at least to the information contained in them. Continued
16
access to digital materials for a dened time (but not indenitely) is mediumterm preservation: here, the time period is long enough to encompass changes in technology. Short-term preservation is, in part, dened by changes in technology: access to digital materials is maintained until technological changes make it inaccessible, or for a period during which the material is likely to be in use but which is relatively short (Jones and Beagrie, 2001, p.10). Such denitions provide helpful ways of thinking about digital preservation programmes: for example, about resource allocation, the long-term resource implications of embarking on long-term preservation being ongoing and therefore large.
What strategies and actions do we need to apply when we preserve them?

The denitions in these two sources make us aware, in a general sense, of the components of a digital preservation programme. In order to achieve the aim of a digital preservation programme maintaining accessibility of digital objects over time (UNESCO, 2003, p.21), or to ensure continued access to digital materials for as long as necessary various processes forming a series of managed activities (Jones and Beagrie, 2001, p.10) are required. They need to form a set of coherent arrangements (UNESCO, 2003, p.21). These are noted in Chapters 7 to 10.
Conclusion
This chapter has introduced some of the key concepts that are reshaping preservation practice in the digital environment. It has noted the need for new ways of thinking about preservation and has posed three key questions that need to be considered when we think about the preservation of digital materials: What exactly are we trying to preserve? How long are we preserving them for? What strategies and actions do we need to apply when we preserve them? These questions and other themes introduced in Chapter 1 are explored in the rest of this book.
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Chapter 2
Why do we Preserve? Who Should do it?
Introduction
Society, of course, has a vital interest in preserving materials that document issues, concerns, ideas, discourse and events . . . The ability of a culture to survive into the future depends on the richness and acuity of its members sense of history (Task Force on Archiving of Digital Information, 1996, p.1) Preservation is commonly perceived to be the responsibility of large, wellresourced institutions such as national libraries and archives, state libraries, and some university libraries. This perception is no longer valid in the digital age. It may once have been legitimate, in the days when conservation laboratories were considered a necessary requirement for a successful preservation programme, and when computer installations were expensive and few in number, but the reality is now very different. Documentary materials in digital form are now being created at all levels of society. Digital preservation must come within the responsibilities of almost every creator and every user of digital information, not just those of librarians and archivists. This chapter investigates the questions: Why should digital materials be preserved? Who has responsibility for their preservation? How signicant is the problem of digital preservation? While the continuing roles of those institutions traditionally identied as responsible for preservation libraries, archives and museums are noted, attention is paid to the roles of the much wider range of stakeholders now required for effective digital preservation.
18
Why preserve digital materials?

The reasons for preserving knowledge are variously described in terms of duties, obligations, and benets. Preservation is based on the notion that, because man learns from the past, evidence of the past therefore has considerable signicance to the human race and is worth saving (Harvey, 1993, p.6). Not only is preservation worth doing; it is also, some suggest, a duty. Agresto, former head of the United States National Endowment for the Humanities, suggested that we have a human obligation not to forget (cited in Harvey, 1993, p.7). The importance of preservation for a healthy society is frequently noted, one example being Agresto again, who considered that preservation is essential for the well-being of democracies which depend on knowledge and the diffusion of knowledge and on knowledge shared (Harvey, 1993, p.7). (Many commentators seem unaware that democracy is not the only possible foundation of a viable society, but we leave aside here the question of the role of preservation in societies not based on democratic principles. There is considerable evidence that those in power in democracies are just as likely to suppress or destroy evidence that does not suit their political agenda as are those in power in societies based on other political systems.) It is also often contended that nothing less than the ability of a culture to survive into the future depends on the preservation of knowledge (Task Force on Archiving of Digital Information, 1996, p.1). The cultural and political imperatives that have led to preservation have been explored in such books as Lowenthals The Past is a Foreign Country (Lowenthal, 1985) and Taylors Cultural Selection (Taylor, 1996), which persuade us that preservation is not simply the concern of a limited number of cultural heritage institutions and professions, but has dimensions that have signicant impact, both limiting and sustaining, on most aspects of society. Economic reasons for preservation are noted by others attuned to the current pragmatic political environment. Brindley, the British Librarys Chief Executive, notes that there are sound economic reasons for preserving the digital assets that are resulting from large investments in digitizing projects (Brindley, 2000, p.125). There is, in fact, no single reason why we preserve knowledge. Preservation, suggests Cloonan, has a life force fueled by many (often disparate) sources (Cloonan, 2001, p.231). None of these reasons change when we consider the preservation of knowledge encoded in digital materials, but the rhetoric alters to emphasize economic rationales. Preserving digital materials is essential if mankind is not to waste the institutional and societal investment already made in digital resources (B. Smith, 2002, p.135); the emerging knowledge economy cannot survive and will not lead to democratic outcomes without a provision for the archiving function (Waters, 1996, p.4). Recordkeeping and accountability reasons are also commonly given: we expect that this [digital] content will remain accessible to allow us to validate claims, trace what we have done, or pass a record to future generations (NSF-DELOS Working Group on Digital Archiving and Preservation, 2003, p.[i]). The NSF-DELOS Working Group on Digital Archiving and Preservation species ve conditions, any one of which, if it exists, is sufcient to provide a benet to society: If unique information objects that are vulnerable and sensitive and therefore subject to risks can be preserved and protected;
Professional imperatives 19
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
If preservation ensures long-term accessibility for researchers and the public; If preservation fosters the accountability of governments and organisations; If there is an economic or societal advantage in re-using information, or If there is a legal requirement to keep it (NSF-DELOS Working Group on Digital Archiving and Preservation, 2003, p.3).
The report of this Working Group provides examples and further discussion. It provides, for instance, the example of repositories of scientic data supporting the trends towards data-driven science to illustrate the condition of ensuring long-term accessibility for researchers, and the economic benets are noted in relation to reduction of costs through re-use of stored data (NSF-DELOS Working Group on Digital Archiving and Preservation, 2003, p.5). There is also a widespread appreciation that the preservation scene is changing in signicant ways. Lyman and Kahle (of Internet Archive fame) note that recordkeeping, which, combined with archival preservation, is the basis of historical memory, was greatly facilitated by print, and institutions such as universities, publishers and museums collect, organize and preserve the historical memory that gives culture continuity and depth. But this is changing: What are, and will be, the social contexts and institutions for preserving digital documents? Indeed, what new kinds of institutions are possible in cyberspace, and what technologies will support them? What kind of new social contexts and institutions should be invented for cyberspace? (Lyman and Kahle, 1998). The very aims of preservation are also being questioned What are we preserving? For whom? And why? (NSF-DELOS Working Group on Digital Archiving and Preservation, 2003, p.2) and the expanded number of stakeholders in the digital age means that different interests must be considered.
Professional imperatives
What do these changes mean for libraries and archives? Has the basis on which their practice is founded altered? At one level, there have been few changes. In the traditional view, libraries are, according to Marcum, then President of the Council on Library and Information Resources, societys stewards of cultural and intellectual resources (Kenney and Stam, 2002, p.v). Preservation is nothing less than the core business of libraries whose collections are maintained for use into the future (Webb, 2000b). For libraries and archives, preservation is about protecting information of enduring value for access by present and future generations (Hedstrom, 1998, p.189) and documentation of human activity (Conway, 1996, p.3). One typical view of the preservation role of libraries is Gormans statement: Libraries have a duty to preserve and make available all the records of humankind. That is a unique burden. No other group of people has ever been as successful in preserving the records of the past and no other group of people has that mission today . . . Let there be no mistake: if we librarians do not rise to the occasion, successive generations will
20
know less and have access to less for the rst time in human history. This is not a challenge from which we can shrink or a mission in which we can fail (Gorman, 1997). The arguments for preservation being a core activity in libraries are traditionally articulated as maintaining the collection for access for a period of time (which varies according to the type of library and the number of years material is required to remain accessible), as making good sense economically by allowing items to be used longer before they wear out, and the just in case argument: It cannot easily be predicted what will be of interest to researchers in the future. Preserving current collections is the best way to serve future users (Adcock, 1998, p.8). Similarly, there has been no fundamental change in the way that archivists consider preservation. Archivists secure understanding of their preservation responsibilities is unaltered. They have typically placed the physical care of their collections at least on a par with, if not at a higher level of importance than, the provision of access to those collections. This physical defence of archives was indeed considered paramount by the British archivist Sir Hilary Jenkinson, who formulated this statement in 1922: The duties of the Archivist . . . are primary and secondary. In the rst place he has to take all possible precautions for the safeguarding of his Archives and for their custody . . . Subject to the discharge of these duties he has in the second place to provide to the best of his ability for the needs of historians and other research workers. But the position of primary and secondary must not be reversed (Jenkinson, 1965, p.15). There is ample evidence in the literature of their profession to support the generalization that archivists have thought more thoroughly about their professional practice and have articulated it more clearly than have librarians. Gilliland-Swetland examines how archival principles provide powerful ways of thinking about the preservation of digital materials. Implementing the archival perspective in the digital environment, she suggests, encompasses Working with information creators to identify requirements for the long-term management of information; Identifying the roles and responsibilities of those who create, manage, provide access to, and preserve information Ensuring the creation and preservation of reliable and authentic materials; Understanding that information can be dynamic in terms of form, accumulation, value attribution, and primary and secondary use; . . . Identifying evidence in materials and addressing the evidential needs of materials and their users through archival appraisal, description, and preservation activities (Gilliland-Swetland, 2000, p.21). Nor should we forget the specic legal reasons for preservation. In the case of archives these reasons are often connected to administrative and political accountability. For some types of libraries, statutory responsibilities require preservation to be their core business. National libraries, for example, have the role of collecting, and safeguarding access to, information published in their
Other stakeholders 21
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
country. In Australia, the National Library of Australia has a statutory responsibility to collect and preserve signicant Australian publications regardless of their format and has a special responsibility for collecting and preserving the published output of Commonwealth government agencies, including online publications (Gatenby, 2002a). These are the traditional preservation responsibilities of libraries and archives, but they are changing rapidly. Commentators agree that the need for preservation will remain, and most suggest that it will become even more urgent; they also agree that the roles of these institutions will be different. Webb, for instance, considers that these roles will require exibility, a willingness to change, proactively seeking a useful role that draws on the expertise and challenges we already have, while developing whatever new expertise and perspectives will be needed. Preservationists will need to be willing to listen and learn, and to form new alliances and partnerships (Webb, 2000b). The preceding sections are reiterated and summarized cogently by Smith: Society has always created objects and records describing its activities, and it has consciously preserved them in a permanent way . . . Cultural institutions are recognised custodians of this collective memory: archives, librar[ies] and museums play a vital role in organizing, preserving and providing access to the cultural, intellectual and historical resources of society. They have established formal preservation programs for traditional materials and they understand how to safeguard both the contextual circumstances and the authenticity and integrity of the objects and information placed in their care . . . It is now evident that the computer has changed forever the way information is created, managed, archives and accessed, and that digital information is now an integral part of our cultural and intellectual heritage. However the institutions that have traditionally been responsible for preserving information now face major technical, organizational, resource, and legal challenges in taking on the preservation of digital holdings (B. Smith, 2002, pp.133134).
Other stakeholders
The new challenges of digital preservation require new participants: we are seeing the privatisation of preservation (Burrows, 2000, p.143). Stakeholders other than those who participated in the old preservation paradigm, whose legal and moral rights must be considered, are becoming signicant players. Others are starting to participate as it becomes increasingly evident that the cultural heritage institutions traditionally charged with preservation responsibility cannot continue to carry this responsibility in the digital age without widening their range of partners in the endeavour. Digital preservation also affects information on long-term genetic research, monitoring global environmental change, locating nuclear waste sites, establishing property rights, storing and authenticating electronic criminal evidence, etc (B. Smith, 2002, p.135); it is also a consumer problem: Individuals are storing an increasing portion of their social and personal memory on digital media with the mistaken belief that this will ensure
22
that those memories will always be available to them for consultation. Today no one is able to provide such guarantees (B. Smith, 2002, p.135). While some of the digital output of individuals is being stored in their own computing environments, a very large amount of it exists publicly on the web. Not only are new kinds of stakeholders claiming an interest or claiming control, but higher levels of collaboration among stakeholders are seen as necessary for digital preservation to be effective. Localized solutions are not perceived as likely to be the most effective; the problem is large enough to warrant a world-wide effort (B. Smith, 2002, p.136) and the preservation of digital material will become essentially a distributed process where traditional demarcations do not apply and one for which an interdisciplinary approach is necessary (Shenton, 2000, p.164). Collaboration is considered more and more as the only way in which viable and sustainable solutions can be developed, as the problems are well beyond the scope of even the largest and most wellresourced single institution. (UNESCO, 2003, Chapter 11 explores collaboration in more detail, and Chapter 9 of this book provides examples of collaborative activities). Who, then, are these new stakeholders? What are their preservation roles likely to be in an increasingly digital environment? An early expression of this was made by the Task Force on Archiving of Digital Information, whose inuential 1996 report set much of the digital preservation agenda for the rest of the decade. The report suggested that intense interactions among the parties with stakes in digital information are providing the opportunity and stimulus for new stakeholders to emerge and add value, and for the relationships and division of labor among existing stakeholders to assume new forms. It proposed two principles, the rst that information creators, providers and owners have initial responsibility for archiving their digital information objects and thereby assuming the long-term preservation of these objects and the second that where this mechanism fails or becomes unworkable, certied digital archives have the right and duty to preserve digital materials (Task Force on Archiving of Digital Information, 1996, pp.1920). The implications of these principles are still being worked through today. Information providers, creators and owners, and certied digital archives are new stakeholders, but there are many others. Hodge and Frangakis note that other stakeholders identied in the literature include disciplines, commercial services, government agencies, passionate individuals, rights holders, beneciaries, funding agencies, and users (Hodge and Frangakis, 2004, p.15); they add trusted third parties to that list (Hodge and Frangakis, 2004, p.20). UNESCOs Charter on the Preservation of the Digital Heritage lists hardware and software developers, creators, publishers, producers, and distributors of digital materials as well as other private sector partners in addition to libraries, archives, museums and other public heritage organizations (UNESCO, 2004, Article 10). Lyman and Kahle suggest that many different communities must learn to work together if the problem of digital preservation and archiving is to be solved computer scientists, librarians and scholars, and policy makers (Lyman and Kahle, 1998). Some of these communities were communicating in the old preservation paradigm, one example being the collaboration of scholars and librarians to identify the core literature in specic discipline areas for microlming and scanning projects (see Gwinn, 1993 for example in agriculture). The extent
Other stakeholders 23
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
and nature of collaborative activity, however, will need to intensify. Howell, an Australian preservation manager, indicates other stakeholders: managers of digital archives, representatives of the public interest and public policy, and actual and potential users of digital information. One major challenge, he suggests, is to meet the requirements of all stakeholders while still ensuring that preservation and access objectives are also met, especially for digital materials which may have commercial value (Howell, 2001, p.140). Another Australian articulation lists the stakeholders in the Australian digital preservation context as government bodies, the corporate and academic sectors, software vendors, creators and publishers, libraries and archives, networked information service providers and the digital material audience of present and future generations (Kerry, 2001, p.13). Some attention has been given to the role of scientists and scholars as stakeholders in digital preservation. The way in which scholars work is being transformed, most notably in the sciences, but also in the social sciences and the humanities, has been investigated. The increasingly data-driven approaches to science that are currently in fashion use very large primary research data sets, particularly in the genomics and earth sciences (Howell, 2001, p.133). This means that increasing attention is being given to the long-term preservation of digital scientic data (see, for example, National Research Council, 1995, ERPANET/CODATA Workshop, 2004; Hodge and Frangakis, 2004). The dissemination of scholarly output provides an example of the transformations of this new-model scholarship (Smith, 2003). No longer do scholars rely solely on formal print publication mechanisms for transmitting their research. Instead, more and more emphasis is placed on other mechanisms: pre-print archives in high-energy physics and in mathematics, e-repositories (often university-based), the development of web sites and internet discussion lists that are based around communities of scholars, all of which have preservation implications. A useful illustration of the major changes that will be required is in the keeping of personal correspondence. Its value as a source of historical information has long been recognized. The extensive use of e-mail has diminished the likelihood that personal correspondence will remain accessible to future historians. Lukesh asks: Where will our understandings of today and, more critically, the next century be, if this rich source of information is no longer available? as scientists, scholars, historians and many others increasingly use e-mail (Lukesh, 1999). Pre-digital paradigm preservation habits mean that scholars, scientists and other creators of e-mails retain the expectation that librarians and archivists will continue to collect and maintain materials ascertained to be of long-term value, usually well after the time of creation of the materials. But with e-mail (and in fact all digital materials) this cannot occur because the digital materials will quickly become inaccessible. (Reasons for this are explained in Chapter 3.) With digital materials, suggests Smith, the critical dependency of preservation on good stewardship begins with the act of creation, and the creator has a decisive role in the longevity of the digital object (Smith, 2003, pp.23). For most creators of digital materials this is a new role. Publishers are another stakeholder group whose responsibilities and roles change as more of their output is distributed in digital form. Some national libraries are developing cooperative arrangements with publishers to ensure that preservation responsibilities for digital publications are understood and shared. One example is the agreement between Elsevier Science and the Koninklijke
24
Bibliotheek (the National Library of the Netherlands) through which the Library receives digital copies of all journals on Elseviers ScienceDirect web platform. The National Library of Australia and the Australian Publishers Association are developing and trialing a Code of Practice for Providing Long-Term Access to Australian Online Publications. This code outlines the conditions and responsibilities that each partner agrees to observe in order to ensure Australian online publications remain available for use into the future (Phillips, 2002). Some express doubts about the interest and willingness of for-prot organizations to participate in digital preservation initiatives. Search engine companies, for instance, are not in the business of long-term archiving of the web or even a portion of it, nor should they be expected to take on this responsibility. The entertainment industry is increasingly digital, and its products, audio and video, have a well-established place as critical resources for research, historical documentaries, and cultural coherence resources. Even given the prevailing political ethos, it is impossible to envisage a situation where market forces will be sufcient to ensure the preservation of this digital material. The opposite may be true: in some cases market forces work against long-term preservation by locking customers into proprietary formats and systems (Workshop on Research Challenges in Digital Archiving and Long-term Preservation, 2003, pp.x-xi). An Australian example illustrates some of the issues and the difculties in nding viable solutions. Pockleys search for a suitable home for his digital compilation The Flight of Ducks is described in his evolving online report Killing the Duck to Keep the Quack (Pockley, 1995- ). This case study makes it clear that some traditional cultural heritage institutions lack knowledge of the issues. Some are unlikely to be able to bear the costs and complexities of moving digital content into the future; others will deliberately or inadvertently, through a simple failure to act, render the information irretrievable; and for some it is easier to ignore the existence of digital work or to treat it as somehow less worthy of collection and preservation (Pockley, 1995-, Section 4. Death in Custody). Three different custodial environments illustrate three sets of problems. RMIT University did not accept custodial responsibility for online research during the development of the work, which was submitted as a doctoral thesis at this university, and it had no infrastructure to support long-term access. Political and ethical issues also intervened, as some members of the Universitys Ethics Committee suspected it might be culturally offensive. The National Library of Australia accepted the site for PANDORA, but has been unable to capture the password-protected area, which contains restricted material, so that the compilation is not preserved in its entirety. Cinemedia, a Melbourne-based organisation established in 1997 to cater for the screen-culture needs in the state of Victoria, next hosted the compilation. However, this arrangement was not fully satisfactory: Cinemedia retained control of all material and had the right to remove the whole or parts at any time, and lacked an infrastructure for preservation of online work (Pockley, 1995- ).
The extent of the preservation problem

It is useful to have some understanding of how substantial the preservation problem is, although estimates must remain very inexact. We know something of its parameters for the traditional artifacts, mainly paper-based, that make up
The extent of the preservation problem 25
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
the collections of libraries and archives. Studies of paper deterioration have been carried out since the 1950s, after Barrows inuential study Deterioration of Book Stock: Causes and Remedies: Two Studies on the Permanence of Book Paper (Barrow, 1959) raised alarm within the library community. Later studies rened these conclusions. Although the evidence allows only general statements to be made, it is commonly thought that paper embrittlement affects in the order of 30 per cent of the collections of large research libraries in the United States. Studies for Australia suggest a lower rate. (This is explained in more detail in Harvey, 1993, pp.910). Nor is the problem of deterioration of the kinds of artifacts typically found in traditional collections limited to paper. All materials deteriorate. Photographs, nitrate lm, and cellulose acetate lm are some examples. However, our primary interest in this book is digital materials. What is the extent of the preservation problem for these? The geometrical increase in the volume of digital materials (Webb, 2000b) means that there is much more digital content worth preserving, and it follows that this quantity is also increasing geometrically, but our ability to manage and preserve them is far outpaced by this growth rate (Workshop on Research Challenges in Digital Archiving and Long-term Preservation, 2003, pp.24). It is suggested the last 25 years have been a scenario of data loss and poor records that has dogged our progress so if this is not reversed, the human record of the early 21st century may be unreadable (Deegan and Tanner, 2002). Such rhetoric is commonly found in the literature, but it is lamentably short on supporting specics and evidence. Alarmist descriptions abound: there will be a digital black hole . . . truly a digital dark age from which information may never reappear (Deegan and Tanner, 2002) if we do not address the problem, and we will become an amnesiac society. (An early use of this evocative term in relation to digital materials was by Sturges (1990), still worth reading today.) Although we may be the best documented era in history much of this has been lost, according to some: the rst email message, chat group session, and web site (Vogt-OConnor, 1999). There are regrettably few documented examples of digital data loss. Usually the literature notes only general categories of digital material that are, or are thought to be, at risk. One of these comes from the US federal government, for whom, OMahony notes, it was the norm, when web sites or internet les were changed, to overwrite the old information. He concludes that the public is now experiencing losses of government information, on a scale similar to that of the catastrophic re of 1921 [in which the 1890 census records were destroyed], on what seems to be a regular basis (OMahony, 1998, p.108). Another example is of business records in electronic form. A study of the companies that were relocated after the World Trade Centre bombing in 1993 found that 40 per cent ceased trading, a major reason being the loss of key business records in electronic form, and 43 per cent of companies which lose their data close down (cited in Ross and Gow, 1999, p.iii). Some of the documented examples are about data recovery, in the sense that data thought to have become inaccessible were able to be recovered, or thought to have been lost were only mislaid. Ross and Gow note four case studies of data recovery: the Challenger Space Shuttle Tapes, Hurricane Marilyn (Virgin Islands September 1995), video image recovery from damaged 8mm recorders (from a crashed ghter plane), and German unication and the recovery of electronic records from East Germany (Ross and Gow, 1999, pp.3942). Another documented example is of the Functional Requirements
26
for Evidence in Recordkeeping project administered by the University of Pittsburgh. Due to a technical glitch at the School the web site with the working les of this project was destroyed, but since the web site has not been updated since 1996 when the Project ended individuals interested in the project and use if the site can access it through the Internet Archive (Cox, 2000?). From the literature it is only possible to conclude, as do Jones and Beagrie, that the evidence of digital data loss is, overall, anecdotal. Jones and Beagrie note, for the United Kingdom, that, while the evidence is as yet only largely anecdotal . . . it is certain that many potentially valuable digital materials have already been lost, and cite the conclusion of the Catriona II project that, although digital resources were being created by Scottish universities, they were largely inaccessible (Jones and Beagrie, 2001, p.18). What is especially disturbing is that all the evidence of loss comes from large institutions, which might be expected to have the knowledge, resources to maintain access to these materials (Tibbo, 2003, pp.1415).
How much data have we lost?

Can we quantify the extent of the preservation problem for digital materials? Why would we want to do this? We need to document some specic examples to supplement the handful of case studies (many of them anecdotal) that are currently all that can be found. Some of these examples are probably present in unpublished reports or business cases, but the publicly-accessible literature contains little. Beagrie suggests that statistics on current losses are difcult to compile. Wider overviews are rare (Beagrie, 2004, pp.12). If we cannot produce specic examples, we lay ourselves open to the charge of being alarmist. A useful question to pose, then, is this: Is the problem of digital preservation as great as we have assumed? One commonly-applied criterion for measuring the quality of a piece of research or scientic investigation is to ascertain that it has referred to (and preferably has built on) what has already been investigated; a literature review is the usual way in which the results of this investigation is presented. Another criterion is the presence of a clear statement of the problem to be investigated. In other words, we expect a clear understanding of the nature and extent of the problem to be demonstrated. Applying these criteria to the preservation of digital information would call for, at the least, a literature search to ascertain, as accurately as possible, just how large the problem is. We need answers to these questions: How much digital information has been lost and how much has been compromised? To what extent has the data been compromised? Is the problem of digital preservation as great as we have assumed? We assume that there is a problem, and many statements, including those in this book, have been based on the assumption that there is a problem. Phrases such as data losses through poor management of digital data (Beagrie and Greenstein, 1998, Preface) and there is general consensus that digital preservation is both an imperative and a challenge (Lavoie, 2003, p.2) are readily found but little else. It is valid, therefore, to ask how large the problem is it may even be crucial to the future of digital preservation, for instance to secure resources. An inability to answer this question could lead to scepticism about whether the problem is as great as claimed and to conjecture that we may have been needlessly
How much data have we lost? 27
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
alarmed into believing the issue is more urgent than it really is. An attempt to answer these questions is especially valid in the context of the history of the preservation of documentary heritage. Two examples brittle books and nitrate lm come to mind. In both of these the initial concerns were undoubtedly wellmeaning and were based on the best information available at that time, but the assumptions on which action was taken have been established as faulty to varying extents, with serious consequences. The preceding section in this chapter noted the extent of the preservation problem for paper-based material. This concern precipitated expenditure of large amounts of money by the US National Endowment for the Humanities on its Brittle Books Program, which has microlmed about one million books since it began in 1988. But was this money well spent? There is now some evidence to suggest that the rate at which paper becomes brittle is not as high as the earlier studies, on which the Brittle Books Program was based, suggested. This has led to some strident critiques (the most notable among them that of Baker (2001)). Similarly, early projections about the rate at which cellulose nitrate lm would deteriorate resulted in large-scale reformatting of these lms to cellulose acetate stock, but there is now considerable concern about the stability of the new archival carriers (ANICA, 2002). The examples of brittle books and nitrate lm might suggest that we urge caution in the case of digital information. This is an excellent reason to put some effort at an early stage into attempting to quantify the extent of digital information loss or compromise, or, at the very least, to document some specic examples to supplement the few studies. The desirability of more documented examples and case studies has been recognized for some years. For example, Ross and Gow concluded that information about data loss, recovery, and risk is very difcult to acquire . . . more case studies about data loss and rescue need to be collected (Ross and Gow, 1999, p.vi) and James noted that documented examples of historically signicant data loss are rare (James, 2001). The term data loss used here also includes data that is compromised: it is degraded to the extent that its quality is affected. (The phrases loss of data integrity or loss of data authenticity might also be used.) The data may still be accessible, but we have no clear idea of what they mean, what software was used to create them, and so on. The question How much digital information has been lost and how much has been compromised? is difcult to answer maybe impossible to answer. Although studies of the kind of data which should be preserved are readily available, such as scientic data (National Research Council, 1995), historical data (Higgs, 1998) and statistical data (Royal Statistical Society and UK Data Archive, 2002), no general estimates of quantity based on solid evidence (as distinct from conjecture) have been located, and few specic examples or carefully documented case studies appear to exist. The same examples are trotted out, even where they are no longer in the lost or compromised category: the BBCs Domesday Project, NASA data, the Viking Mars mission, the Combat Area Casualty le containing prisoner of war and missing in action information for the Vietnam war, the rst e-mail, the rst web site, as described in more detail below. It is doubtful that we will be able to accurately quantify the extent of loss. If we could, the answer would be inextricably bound up with the issue of selection for preservation. Are we assuming that we need to preserve most, perhaps even all, digital materials, rather than be selective? Do we have unrealistic
28
expectations about the quantity of information (and not just digital information) that we want to preserve? As a crude (and unlikely) example, if we assume that all e-mails need to be kept, then the loss of any becomes a problem. One argument is that anything signicant is likely to be maintained anyway, so should we be concerned about the rest? Some of the examples which follow assume that the rst (e-mail, web site, and so on) is worth preserving but is this necessarily the case? It is, often, a view developed in hindsight. Betts tells us that Ray Tomlinson, principal engineer at BBN Technologies in Cambridge, Massachusetts did not save the rst network e-mail ever sent in 1972 because it just didnt seem worth saving . . . Even if backup tapes did exist, they might not be readable. They were just mag tapes, and after seven or eight years, the oxide starts falling off, especially from tapes of that era (Betts, 1999). Selection is noted in more detail in Chapter 4. Specic examples, although few in number, provide an indication of how great the problem of loss or compromise of digital materials may be. The most often quoted, indeed overused, examples are those cited in the 1996 report of the Task Force on Archiving of Digital Information. Because they have been reported very widely since, they warrant quoting at some length. The report notes the case of the US Census of 1960. As it compiled the decennial census in the early sixties, the Census Bureau retained records for its own use in what it regarded as permanent storage. In 1976, the National Archives identied seven series of aggregated data from the 1960 Census as having long-term historical value. A large portion of the selected records, however, resided on tapes that the Bureau could read only with a UNIVAC type-II-A tape drive. By the mid-seventies, that particular tape drive was long obsolete, and the Census Bureau faced a signicant engineering challenge in preserving the data from the UNIVAC type II-A tapes. By 1979, the Bureau had successfully copied onto industry-standard tapes nearly all the data judged then to have long-term value. The report notes the effect of this well-publicized loss: the data rescue effort was a signal event that helped move the Committee on the Records of Government six years later to proclaim that the United States is in danger of losing its memory. The Committee did not bother to describe the actual details of the migration of the 1960 census records. Nor did it analyze the effects on the integrity of the constitutionally-mandated census of the nearly 10 000 (of approximately 1.5 million) records of aggregated data that the rescue effort did not successfully recover. Instead, it chose to register its warning on the dangers of machine obsolescence in apocryphal terms. With more than a little hyperbole, it wrote that when the computer tapes containing the raw data from the 1960 federal census came to the attention of NARS [the National Archives and Records Service], there were only two machines in the world capable of reading those tapes: one in Japan, and the other already deposited in the Smithsonian as a relic . . . Other examples . . . equally illustrate how readily we can lose our heritage in electronic form when the custodian makes no plan for
How much data have we lost? 29
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
long-term retention in a changing technical environment. In 1964, the rst electronic email message was sent either from the Massachusetts Institute of Technology, the Carnegie Institute of Technology or Cambridge University. The message does not survive, however, and so there is no documentary record to determine which group sent the pathbreaking message. Satellite observations of Brazil in the 1970s, crucial for establishing a time-line of changes in the Amazon basin, are also lost on the now obsolete tapes to which they were written (Task Force on Archiving of Digital Information, 1996, pp.23). Rothenberg reminds us of some examples noted in a 1990 US House of Representatives report: hundreds of reels of tape from the Department of Health and Human Services; les from the National Commission on Marijuana and Drug Abuse, the Public Land Law Review Commission, the Presidents Commission on School Finance, and the National Commission on Consumer Finance; the Combat Area Casualty le containing POW and MIA information for the Vietnam war; herbicide information needed to analyze the impact of Agent Orange; and many others (Rothenberg, 1999b, pp.12). He reiterates the paucity of specic examples and offers a reason: To date there appear to be few documented cases of unequivocal loss, but this may simply reect the fact that documents or data that are recognized as important while they are still retrievable are the ones most likely to be preserved (Rothenberg, 1999b, p.2). Another example frequently cited is the BBC Domesday Project. This project captured the national imagination in the UK and resulted in a multi-media version of the Domesday Book on videodisc, produced to mark the 900th anniversary of the original. It became inaccessible in the late 1980s as the hardware platform for which is was developed, the BBC microcomputer, became obsolete. The data has since been restored using emulation techniques (Digital Domesday Book Unlocked, 2002; CAMiLEON, 2002?; Mellor, 2003). This project is also noted in Chapter 7. Cooks 1995 call to action gives some Canadian examples of data loss where recordkeeping practices were ignored in the move to online recordkeeping. Cook noted that the National Archives of Canada . . . found not only that 30 out of 100 randomly chosen policy documents could not be found in the governments paper records, but also that no system was in place to safeguard the contents of the electronic system. Ontario Hydros nuclear power plant failed to keep adequate electronic or paper records of its construction and operation (Cook, 1995). Some disciplines and industries have investigated their own practices and here the evidence is somewhat rmer. One of these is archaeology in the United Kingdom. The growth in computer use and the increase in digital data sets used by archaeologists resulted in the establishment of the Archaeology Data Service, part of the AHDS (Arts and Humanities Data Service), whose web site provides
30
studies. One such study is of the Newham Museum Archaeological Service archives, described here as an indication both of the extent of data loss and of the reasons why data is lost (Dunning, 2001). The Newham Museum Archaeological Service closed suddenly in 1998. Data on the hard disks of its computers were transferred onto oppy disks and were presented to the Archaeology Data Service (ADS) in the hope that the data could be retrieved. There were 239 disks with a total of 6350 les containing graphics, text and data in a variety of formats. Twenty-ve les were corrupted and 12 were unrecoverable. However, the lack of documentation made much that was recovered unusable for archaeological purposes because the data made no sense. This case study teaches us that The loss of data from the Newham Archive is as much to do with poor project planning as it is to do with preservation . . . Preserving and documenting data should not be an additional chore at the end of a project, but an on-going process that is integrated into the creation of the data (Dunning, 2001). Scientic data have been the focus of many studies. One is Preserving Scientic Data on Our Physical Universe (National Research Council, 1995), which indicates what scientic data were then available from United States scientic observation and what they have been and might be used for. It includes some comments about what has survived. Space physics data are among these. This area of research has generated about 50 gigabytes of data per year over the last 30 years and much of this was archived by sending the tapes, also sometimes relevant documentation, to the NSSDC (National Space Science Data Center). However, there are many data at the NSSDC that most scientists would nd difcult to use with only the information originally supplied (National Research Council, 1995, p.21). This report also notes the Landsat data, a large part of which resided on tapes that cannot be read by any existing hardware. Recent data-rescue efforts have been successful in getting older data into accessible form, but these efforts are time-consuming and costly (National Research Council, 1995). Humphrey gives examples of the research data generated by research funded by the Social Sciences and Humanities Research Council of Canada. Of a set of 150 studies from 1977 to 1980, only datasets from three could be located in 1998 (Humphrey, 2003). Other examples come from a range of areas. For electronic journals, Warner notes the irony of the demise of an archive established to archive electronic journals due to lack of funding. The CICNet Journal Archive archived electronic journals from 1991 to 1997 but, according to Warner, has now vanished (Warner, 2002, quoting Wiggins, 2001). Warner also notes the response to a query from Science Direct [Elseviers web platform] was that at least 2% of its electronic journal content is missing (Warner, 2002). Commercial data recovery companies provide snapshots. The web site of Authentec International provides six case studies of loss of data, much of which were subsequently recovered: for example, 15 per cent of the data on two diskettes chewed by a dog were not recoverable, and for a laptop computer damaged in the cargo hold of a plane 36 out of 112 bad sectors of its hard drive were unrecoverable (Authentec International, 2003). dataVault, citing research from CBL, Gartner Research, Meta Group, and SunGard Data Systems, reports that 94 per cent of all PC users surveyed experienced at least one signicant data loss, on average, once a year, that fewer than one per cent
Current state of awareness of the digital preservation problem 31
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
of companies perform daily data backups, and only 20 per cent of those that back up their data store backups offsite (Global Data Vault Inc., 2005). Although the web site of Ontrack Services does not provide specic examples of data loss, it does indicate the high costs of data recovery and the causes of data loss (Ontrack Data Recovery, 2004). While studies of document persistence on the web (such as Koehler, 2004) are primarily studies of access to data, they also throw some light on how much data is lost. Smith examined the longevity of a number of Australian winery web sites and concluded that, although a high percentage of them were still accessible, there was a signicant loss of data within the sites (Smith, 2004). That the Electronic Literature Organization has mounted a project to identify threatened and endangered electronic literature and promote its protection (Electronic Literature Organization, 2005) is another example of action motivated by the assumption that digital materials have been lost, or are, at the very least, are threatened with loss.
Current state of awareness of the digital preservation problem

One cause of data loss is lack of awareness about the need for proactive thinking and interventionist action if we hope to preserve digital materials. What is the current state of awareness of the problem? Wheatley describes the lack of understanding by IT industry vendors as demonstrated during the BBC Domesday Project. Commercial vendors offered to assist this digital preservation project: [one] offered us a special polymer that they guaranteed would preserve a CD-ROM for 100 years. They were unable to answer how we would preserve a CD-ROM player for that length of time. Wheatley concludes that the myth that long-lived media equals long-lived preservation is still worryingly popular (Abbott, 2003, p.10). The general consensus is that levels of awareness of digital preservation are not keeping pace with the expansion in digital resource creation. There is general awareness of the problem but not much beyond that, except in a few isolated discipline areas, such as geoscience, where considerable specialist expertise is to be found. This is related to a continuing lack of appreciation by information professionals that preservation of any materials at all should be a central professional concern. Marcum, in describing the awareness of preservation in American academic libraries over the last two decades, notes little change, apart from the increasing specialization of preservation activities: Preservation treatments are increasingly sophisticated and effective, yet preservation as a core activity of libraries remains less visible than others such as cataloging and users surveys (Marcum, in Kenney and Stam, 2002, p.v). Views of the current situation, especially in regard to digital preservation, corroborate Marcums statement. Lukesh notes the apparent hole in the literature
32
on the subject which seems, she suggests, not to have risen to a level of importance for historians, biographers, librarians and archivists (Lukesh, 1999). It remains too often the concern of only a small number of information professionals, as Brindley, the British Librarys Chief Executive, pointed out during her keynote address to a digital preservation conference in 2000: digital preservation is quite clearly not attention grabbing enough . . . to have yet brought seriously on board authors, publishers and other digital content creators, funding agencies, senior administrators, hardware and software manufacturers, and so on. Take a look at the conference attendance list if you do not believe me (Brindley, 2000, p.127). Others are more sanguine. The level of awareness is increasing, for instance among scholars, who were keen to do the right thing but frequently lacked the clear guidance and institutional backing to give them condence about what they should do. It continues to be low among funding agencies and senior administrators who set policy and strategic directions. More guidance is needed for a wide range of people with varying levels of awareness about and expertise of digital preservation (Jones and Beagrie, 2001, p.3). Staff members of US college and research libraries are aware of the issues in general terms, but lack specic knowledge about how to address it. Their levels of awareness form a spectrum; at one end are those who are only beginning to appreciate the impact of digital preservation at the local level, and at the other are those who are taking concrete, if tentative, steps to meet the challenge (Kenney and Stam, 2002, p.9). These low levels of awareness pose a continuing threat (NSF-DELOS Working Group on Digital Archiving and Preservation, 2003, p.i). This point is specically noted in Articles 3 and 4 of the UNESCO Charter on the Preservation of the Digital Heritage. Attitudinal change has fallen behind technological change with the consequence that the threat to the economic, social, intellectual and cultural potential of the heritage the building blocks of the future has not been grasped (UNESCO, 2004, Article 3). To address this threat awareness-raising and advocacy . . . alerting policy makers and sensitizing the general public to both the potential of the digital media and the practicalities of preservation are urgent (UNESCO, 2004, Article 4). In an Australian example discussing digital preservation at the Australian Centre for the Moving Image, Pockley notes, in relation to preservation metadata, that our stakeholders [individuals and industry] do not understand the need for authoritative metadata; have deeply entrenched paper-based or industry work practices; are uncertain about the ownership of information; are not used to working and thinking in an electronic environment. Collaboration is slow because institutions have failed to raise understanding beyond a conceptual level (Pockley, 2002). Although this example relates specifically to creating metadata, it certainly applies to other aspects of digital preservation.
Conclusion 33
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Conclusion
This chapter has dwelt on the question of who should take responsibility for digital preservation. As well as the organizations who have traditionally been concerned with preservation (libraries, archives, museums), the preservation of digital materials must involve many other stakeholders. Their input is required to decide what is kept, negotiate rights management, locate and manage the resources and develop the procedures and policies (Kerry, 2001, p.17). How this is to be done is not yet clear. This chapter has also considered the question of how much digital material we have lost. Although the parameters of loss are very unclear, there is little doubt that the amount of digital materials that we are unable to access, or able to access only after considerable effort and expense, is signicant and will continue to increase unless action is taken. Our current ability to do this is hampered by the widespread lack of awareness of the problem. The next chapter looks more closely at why there is a problem by considering the nature of digital materials.
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Chapter 3
Why Theres a Problem: Digital Artifacts and Digital Objects
Introduction
Old bit streams never die they just become unreadable (Rothenberg, 1999b, p.2) On the surface, digital technology appears to offer few preservation problems. Bits and bytes are easy to copy, so there should be no problems in developing an unending chain of copies into the future, and having copies all over the world in case of disaster. However, we already know that the reality is not so simple and that there are very signicant technical and management problems. The two main factors leading to inaccessibility of digital information: changing technology platforms and media instability, are relentless, with the potential to render digital information useless (Webb, 2000b). Why are digital materials different? What are the modes of digital death? The answers to these questions (posed in Jones and Beagrie, 2001, p.19 and Wiggins, 2001) provide a helpful framework for understanding why digital preservation poses major challenges. In essence there are three sets of challenges: those relating to the nature of the media that are used to store digital materials; those resulting from the technologies required to create, store and access digital materials those characterized in this book as challenges to the integrity of digital materials. The issues are complex and are interrelated (Rothenberg, 1999a, p.1). Rothenbergs inuential statement about the issues is worth quoting at length, as it neatly describes the range of issues and their close relationship:
36
It is now generally recognized that the physical lifetimes of digital storage media are often surprisingly short, requiring information to be refreshed by copying it onto new media with disturbing frequency. Moreover, most digital documents and artifacts exist only in encoded form, requiring specic software to bring their bit streams to life and make them truly usable; as these programs (or the hardware/software environments in which they run) become obsolete, the digital documents that depend on them become unreadable held hostage to their own encoding. This problem is paradoxical, given the fact that digital documents can be copied perfectly, which is often naively taken to mean that they are eternal . . . In addition to the technical aspects of this problem, there are administrative, procedural, organizational, and policy issues surrounding the management of digital material. Digital documents are signicantly different from traditional paper documents in ways that have signicant implications for the means by which they are generated, captured, transmitted, stored, maintained, accessed, and managed . . . [mandating] new approaches to accessioning and saving digital documents to avoid their loss. These approaches raise nontechnical issues concerning jurisdiction, funding, responsibility for successive phases of the digital document life cycle, and the development of policies requiring adherence to standard techniques and practices to prevent the loss of digital information. However, few of these nontechnical issues can be meaningfully addressed in the absence of a sound, accepted technical solution to the digital longevity problem (Rothenberg, 1999a, p.2). Current trends in information management and delivery militate against the effective preservation of digital materials. The sheer number of digital materials now in existence threatens to overwhelm and the rate at which they are being created is increasing (as noted in Chapter 1). The ownership of intellectual property rights in digital materials is often complex (consider, for instance, a lm with scriptwriters, actors, directors and many other interests) and these rights need to be negotiated before libraries and archives can legally preserve some digital materials. New formats continue to proliferate. Many libraries are focusing their attention more on short-term goals with the consequence that collection maintenance over the longer term is not adequately resourced (Webb, 2000b). This chapter is primarily concerned with the reasons why digital materials are difcult to maintain access to into the future that is, to preserve. It examines the causes of deterioration of the media, noting both structural and manufacturing causes and those relating to the storage and handling of the media. It considers the reasons for technological obsolescence and its consequences, such as the loss of functionality of access devices, manipulation and presentation capabilities, and contextual information. This chapter does not include strategies and techniques used to address the deterioration of digital materials that is the focus of Chapters 6, 7 and 8 but they are mentioned in passing. Case Study 1, in the Appendix, examines digital storage at the National Film and Sound Archive and illustrates the topic of media longevity.
Modes of digital death 37
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Modes of digital death

Jones and Beagrie compare the challenges involved in the preservation of digital materials maintaining access to digital materials over time with preservation of paper-based materials. First, they note the implications of the technologyrelated challenges the dependence on machinery for playback of digital materials and the speed at which this technology changes which call for more immediate preservation action than paper-based materials, for which a period of inaction is not usually harmful. They point out that preservation of digital materials cannot be sporadic: a continual programme of active management is needed from the design and creation stage, and this, in turn, results in changing roles of stakeholders and in different kinds of inter- and intra-institutional collaboration. Secondly, the issues resulting from the fragility of the media are noted and, thirdly, the challenges of ensuring the continued integrity, authenticity, and history of digital materials, which are a consequence of the ease of their alteration. Jones and Beagrie also describe some of the implications arising from the nature of digital materials: if proactive preservation is not implemented at an early stage in the life of the digital materials, they will be lost or unusable in a short period of time. This has signicant implications for allocation of preservation resources (Jones and Beagrie, 2001, p.19). An Australian statement of the challenges of digital preservation amplies Jones and Beagries statement. The National Library of Australias Digital Preservation Policy (NLA, 2002) suggests that the challenges fall into three categories: technology-related challenges, media-related issues, and the challenges of ensuring integrity and authenticity. The technology-related challenges include rapid changes in the availability of hardware, software and other technology required for access. The media-related issues are widespread use of relatively unstable media and the diverse and frequently changing range of le formats and standards. The challenges of ensuring integrity and authenticity include uncertainty about the signicant properties that must be maintained for different digital resources and new models of ownership that impose property and other rights-based constraints. The National Library of Australias policy also pays attention to some of the more general issues that have an impact on the preservation of digital materials, such as their large quantity and rapid growth-rate, the recurrent, critical and demanding nature of the threats to accessibility, and uncertainty about viable strategies and techniques to apply. Other challenges and uncertainties are the resource implications (the high costs of taking action, administrative complexities in ensuring timely and cost-effective action is taken over very long periods of time) and the roles of many stakeholders, which are, as yet, ill-dened. Examples to illustrate these challenges are readily located (some in Chapter 2). Ontrack Data Recovery, a large US-based data recovery company, suggests that hardware and system problems provide by far the largest number of reasons for customers to approach them (56 per cent), followed by human error (26 per cent). Software corruption or program problems, computer viruses and natural disasters are the other reasons listed (Ontrack, 2003). Hedstrom and Montgomerys 1998 survey of Research Libraries Group members requested details of current storage medium and format, the date at which the earliest digital materials were written to their current storage medium, and whether
38
they held digital materials that could not be accessed. Their survey reported that the oldest digital materials were written to their current carrier in the late 1970s, although most carriers were only one to three years old. Fifteen of the 36 institutions responding to the survey could not access some digital materials they held because they lacked the operational and/or technical capacity to mount, read, and access them. This was especially the case with oppy disks and open-reel nine-track tape, and to a lesser degree with CD-ROMs, magnetooptical disks, DDS DAT tape, 3480 cartridges, and various audio and video formats (Hedstrom and Montgomery, 1999, pp.1112). Figure 3.1 lists the threats to digital continuity (continuity of production, continuity of survival, continuity of access) identied in the UNESCO Guidelines. While not all of these threats are specic to digital materials, the list serves as a useful reminder of the magnitude of the challenge that we face in preserving digital materials.
The carriers used to store these digital materials are usually unstable and deteriorate within a few years or decades at most Use of digital materials depends on means of access that work in particular ways: often complex combinations of tools including hardware and software, which typically become obsolete within a few years and are replaced with new tools that work differently Materials may be lost in the event of disasters such as re, ood, equipment failure, or virus or direct attack that disables stored data and operating systems Access barriers such as password protection, encryption, security devices, or hard-coded access paths may prevent ongoing access beyond the very limited circumstances for which they were designed The value of the material may not be recognised before it is lost or changed No one may take responsibility for the material even though its value is recognised Those taking responsibility may not have adequate knowledge or facilities There may be insufcient resources available to sustain preservation action over the required period It may not be possible to negotiate legal permissions needed for preservation There may not be the time or skills available to respond quickly enough to a sudden and large change in technology The digital materials may be well protected but so poorly identied and described that potential users cannot nd them So much contextual information may be lost that the materials themselves are unintelligible or not trusted even when they can be accessed Critical aspects of functionality, such as formatting of documents or the rules by which databases operate, may not be recognised and may be discarded or damaged in preservation processing.
Figure 3.1 Threats to Digital Continuity (From UNESCO, 2003, p.32)
Digital artifacts 39
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Digital artifacts
The term digital artifact is here used to mean the carrier, or medium, on which digital data is recorded. Common examples are diskettes, CDs and magnetic tapes. The key issue with keeping digital artifacts can be concisely stated: digital materials are especially vulnerable to loss and destruction because they are stored on fragile magnetic and optical media that deteriorate rapidly and that can fail suddenly from exposure to heat, humidity, airborne contaminants, or faulty reading and writing devices (Hedstrom and Montgomery, 1999, p.1). Digital artifacts are Thibodeaus physical objects, simply an inscription of signs on a medium whose preservation is necessary but not sufcient for effective preservation of digital materials (Thibodeau, 2002, pp.67). Rothenberg brought the issues associated with the short life-span of digital artifacts to public attention in his 1995 article in Scientic American (Rothenberg, 1995, expanded in Rothenberg, 1999b). He pointed out that the physical lifetimes of digital artifacts are often surprisingly short (Rothenberg, 1999a, p.2) but that the actual life-span was irrelevant because it usually exceeded the length of time during which the media could be read. This is a consequence of the rapid obsolescence of hardware and software in a computing environment driven by rapid decreases in storage density and computing costs, and rapid increases in computer speeds. His conservative estimates (Rothenberg, 1995) were: Magnetic tape Videotape Magnetic disk Optical disk 1 year 12 years 510 years 30 years equipment equipment equipment equipment obsolescence obsolescence obsolescence obsolescence 5 years 5 years 5 years 10 years.
A more recent comparison of data carriers is provided in the UNESCO Guidelines, noted in Figure 3.2.
Carrier Current storage capacities per unit Speed of increase in capacity Expected usable life of single unit Other comments generally xed media portable media suitable for backup portable media. unit costs low; low-cost consumer equipment widely available
Magnetic disk up to 200 (eg hard disk) gigabytes Magnetic tape up to 200 gigabytes Optical disk (CD, DVD) up to 4 gigabytes
doubling every around 1218 months 5 years doubling every around 1218 months 5 years slow because not used for very large archives or backups wide range from say, 5 years for low quality products to several decades for high quality products
Figure 3.2 Comparison of Data Carriers (From UNESCO, 2003, p.117)
40
Current advertising in the popular press suggests differently. The readers of popular computing magazines or the computer supplements of daily newspapers are still likely to believe that their digital data will be safe and accessible in the future if it is recorded onto archival quality CD. But the reality is otherwise. The short lifetimes of eight-inch oppy disks, tape cartridges and reels, hard-sectored disks, and seven-track tapes . . . demonstrate, Rothenberg suggests, how quickly storage formats become inaccessible (Rothenberg, 1999a, p.7). One implication of this short life-span is that digital information must be refreshed with disturbing frequency (Rothenberg, 1999a, p.2). The large quantities of materials that are to be preserved exacerbate the challenges. Bradley provides an estimate of the quantity of sound recordings on analogue magnetic tapes that are considered to be of long-term value. These analogue tapes have been, or very likely will be, transferred to a digital medium. It is estimated that there are at least 20 million hours of tapes in European broadcast archives, and perhaps 100 million hours in broadcast collections internationally. There is likely to be as much again in universities and research institutions (Bradley, 2003). Bradley has described in detail the ways in which sound archivists have had to come to terms with both the quantity of digital materials and changing technologies, providing the example from the National Library of Australia, which over ve years moved from analogue tapes to a digital mass storage system for preservation of its sound recordings (Bradley, 2003). Howell suggests six factors that contribute to the rapid deterioration of digital artifacts: manufacturing quality of the medium; how heavily the medium is used; how carefully it is handled; the temperature and humidity levels at which it is stored; the quality of its storage environment; and the quality of the equipment used to access the medium (Howell, 2001, p.139). These are worth examining in more detail. The manufacturing quality of the medium is a key factor simply because all materials deteriorate. In the words of David Bearman, it is a fact of physics, a fact that we must accept as a major limitation on digital preservation, and the outside boundary beyond which we cannot rationally plan to retain the information without transforming the medium (Bearman, 1998, p.24). (We can, however, attempt to inuence manufacturers so that they improve the quality of their product to meet archival and preservation requirements). Knowledge of the physical and chemical makeup of digital artifacts and of the processes of their deterioration is helpful in making decisions about migration. One of the keys to prolonging the life of digital artifacts is the provision of storage conditions that will slow down the rate of deterioration. Howell makes the point that many digital artifacts are rotating technologies (Howell, 2001, p.138). They have moving parts and are, therefore, subject to wear, which can damage the media. For example, even a small rearrangement of the magnetized particles on a magnetic tape, perhaps caused by accidental physical contact with a part of the playback equipment, can result in loss of data; this is often sufcient to corrupt the whole le. The maintenance of the recording and playback equipment to a high standard minimizes the likelihood of this occurring. Similarly, inappropriate handling of digital artifacts can result in physical contact with the part of the media that records the data. An example is touching the surface of a CD-ROM and leaving an oily residue which can corrupt the les stored on it.
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
One of the signicant changes from the pre-digital preservation paradigm is the realization that digital artifacts have little or no artifactual value. This point is explored in the 2001 report of the Task Force on the Artifact in Library Collections. Artifacts are valued in library and archives collections because their physical form demonstrates the originality, faithfulness (or authenticity), xity, and stability of the content (Task Force on the Artifact in Library Collections, 2001, p.vi); the artifact is signicant for research purposes because it provides this evidence. When the information is reformatted, as in microlming a book printed on brittle paper or copying digital data from an obsolete format to a current one, these evidentiary qualities are lost. Because deterioration of digital materials is, as already noted, a fact of physics, which makes migration a fact of digital preservation life, the artifact itself cannot demonstrate qualities such as originality, authenticity or xity. Other mechanisms are required to demonstrate these evidentiary qualities. This is a major change from rmly entrenched pre-digital paradigm thinking, and we are only slowly changing our professional mindsets to accommodate the necessary changes in practice. The following examples of magnetic tape and optical disks illustrate many of the factors that contribute to the deterioration of digital artifacts. Other media types in current use, such as magneto-optical disks and hard disks, and emerging media types, such as HD-ROSETTA etched metal disks, could have been noted. However, the general principles that are identied for the two media types described also apply to these.
Magnetic media
This section is based on the writings of Van Bogart (1995), Ross and Gow (1999), and the International Association of Sound and Audiovisual Archives Technical Committee (2004), which can be referred to for more detailed information. Magnetic tapes have been used for digital data storage from the 1960s. Magnetic media (tape in reels and in housings such as cassettes and cartridges, and disks) are in common use today because they are versatile and cheap, and can provide higher data densities than other media. They are available in a large number of formats: the IASA guidelines for digital audio objects provide the specications of 23 common data tape formats (International Association of Sound and Audiovisual Archives Technical Committee, 2004, p.58), but the number is higher, especially if formats not in current use are also counted. Magnetic tapes store information in the alignment of magnetic particles suspended within a polymer binder, which sticks the magnetic informationcarrying layer to a substrate and provides a smooth surface to assist the tape to run through playback equipment smoothly. Protection of this structure determines the longevity of the tape. If humidity levels are too high, the binder softens or becomes brittle through hydrolysis, resulting in the sticky tape phenomenon where the binder sticks to the equipments tape heads. Data loss from dropout is one result of sticky tape. The magnetic particles, which store data in the direction of the magnetism in them, vary in their magnetically stability. The substrate is usually made of chemically stable polyester lm (Mylar, polyethylene terephthalate, or PET). Mechanical problems affect the substrate, such as stresses on the tape caused by uctuations in temperature and humidity levels in storage areas, which result in mistracking during playback. The substrate can also be stretched if the tape is not appropriately stressed when it is wound or
42
rewound. Other factors that affect data loss include the quality and maintenance of tape recording and playback devices. The longevity of magnetic tape can be improved by attending to its care and handling. Appropriate storage is essential for minimizing deterioration through binder hydrolysis, which is a result of excessive moisture content of the tape. The rate of hydrolysis can be reduced by lowering humidity levels and temperatures in tape storage areas. Magnetic pigments also degrade more slowly at lower temperatures. Also important is that temperature and humidity levels are kept constant and stable. Storage at temperatures that are too high (above 23oC, suggests Van Bogart) increases dropout because the tightness of the tape packing is increased; this, in turn, increases tape distortion. Increased tape-pack stresses also occur as the tape absorbs moisture and expands at relative humidity levels greater than about 70 per cent. Temperature and relative humidity levels that are too high also promote fungal growth. Attention should also be paid to maintaining good air quality, and to reducing dust and debris. Conditioning (acclimatization) is also required if tape is stored in an environment that differs from that in which it is used. (Further information about the care and handling of magnetic media is available in Chapter 7 and in Van Bogart, 1995.) An indication of the life expectancy of some common data tapes is provided in Figure 3.3. The important point to note here is the effect of different relative humidity (RH) and temperature levels on the life expectancy of digital artifacts. A study carried out for the U.S. National Archives and Records Administration (NARA) examined the stability and life expectancy of three high-density magnetic tape types. All three demonstrated an impressive ability to deliver reliable data. Their projected life expectancies were determined to be well outside NARAs requirements. However, a word of caution was sounded: all products suffered some failure during the study. The report notes: As good as the technology is today, failures can and do occur (Arkival Technology Corporation, 2002, p.4). Manufacturers of digital media have concerns that differ from those concerned with digital preservation. Manufacturers are primarily motivated by producing quantity at the lowest cost in high-volume, consumer-driven markets, and by providing new, improved products on a regular basis. The rate of change is rapid: the IASA guidelines for digital audio objects note that all of the main data tape formats have development roadmaps projecting upgrades every 18 months to 2 years (International Association of Sound and Audiovisual Archives Technical Committee, 2004, pp.5455).
Medium D3 magnetic tape DLT magnetic tape cartridge CD/DVD CD-ROM
25% RH 10C 50 years 75 years 75 years 30 years
50% RH 25C
50% RH 28C
3 years
3 years 10 years 9 months
1 year
1 year 2 years 3 months
Figure 3.3 Sample Generic Figures for Lifetimes of Media

(From Jones and Beagrie, 2001, Figure 7)
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Optical disks
This section is based on the work Saffady (1993), Lieberman (1995), Byers (2003), Ross and Gow (1999), and the International Association of Sound and Audiovisual Archives Technical Committee (2004), which can be referred to for more detailed information. The term optical disk is applied to a large number of media which share the characteristic of using laser light to record and retrieve bits from a data layer. Optical disks became available at the end of the 1970s, so we have over three decades of practical experience to draw upon. The IASA Guidelines on the Production and Preservation of Digital Audio Objects note 12 commercially available CD and DVD disk types with storage capacities ranging from 650 megabytes to 9.4 gigabytes (International Association of Sound and Audiovisual Archives Technical Committee, 2004, p.42). All optical disks are structured in basically the same way, but differ to some extent in the way the data is recorded (Lieberman, 1995; Saffady, 1993). Readonly optical disks are produced by a laser which burns pits into a coating on a master disk. From this, another master is produced which is then used to stamp the disks on a plastic base or substrate. The substrate is coated with a thin layer of metal, and is then covered with a protective polymer layer. Recordable optical disks are of two types: record-once (CD-R, DVD-R), and rewritable (CD-RW, DVD-RW). These use different methods for recording data on the disk. Recordonce disks use a dye process, in which the laser light alters the reectivity of the dye layer so that it is read as either reective or non-reective. Different kinds of dyes are used: cyanine (blue); phtalocyanine green); and azo (dark blue). Rewritable optical disks use a phase change (crystallization) process where the recording material (metal alloy lm) is heated. Their archival qualities are not clear: no trustworthy analysis of the medium or long-term reliability has been undertaken (International Association of Sound and Audiovisual Archives Technical Committee, 2004, pp.6566). (Byers provides more detail about the structures of optical disks (2003, pp.512).) The part of an optical disk that is most susceptible to deterioration is the metal reecting layer, because the metal (usually aluminium in CD-ROM and CD-RW disks) is vulnerable to oxidization. Oxidization causes corrosion which obscures the distinction between pit and surface (that is, between 0 and 1, as digital data is stored) and data loss occurs. Metals or alloys less likely to oxidize (platinum or gold, for instance) are sometimes used, but their manufacturing costs are higher. The polymer base supporting the metal substrate may not be impermeable and may encourage oxidization by allowing oxygen to reach the substrate. This polymer base may also promote oxidization in small areas where rough spots or other defects do not provide full protection. Another cause of oxidization is failure of the protective polymer coating. It has also been reported that the ink used to print onto the disk causes the polymer coating to break down, leading to oxidization of the metal layer. For CD-R and DVD-R dyebased disks the density of the dye layer, the dye type, and the amount of exposure to light have an effect on life expectancy and error rates (Kunej, 2004, Slattery et al., 2004). As with magnetic tapes, the longevity of optical disks can be extended by improvements in manufacturing quality and attention to appropriate storage and handling. Although the temporary unreadability caused by ngerprints on
44
a CD-ROM is common knowledge, it is commonly assumed that optical disks are less vulnerable to damage caused by poor handling than are magnetic tapes. This is not the case: all digital artifacts need careful handling, and optical disks are no exception. The protective coating on a CD is thin and care must be taken to ensure that it does not break down. Storage at extremes of temperature and humidity can affect the physical structures: for example, because plastic substrates can absorb moisture, with oxidization of the metal layer following as a consequence, high humidity conditions should be avoided. For the same reasons should be storage areas in which temperature and humidity uctuates, resulting in condensation, should be avoided. (Storage and handling are covered in Chapter 7.) Early studies of optical disk longevity, such as that done at NARA in 1992, suggested that CD-ROMs would last for three to ve years. Consequently NARA did not consider CD-ROM to be an acceptable medium for archival storage, although it was acceptable for use as a transfer medium for permanent records (Harvey, 1995). A 1992 report on accelerated aging studies of CDs, carried out by 3M with input from the National Media Laboratory, suggested a 25-year warranty that assures 100 year life-time at room temperature: that is, the lower estimate takes account of general storage uctuation, as long as its noncondensing, with 100 years as the lifetime in high quality storage conditions (Arps, 1993, pp.102103). Saffady in 1993 summarized manufacturers lifetime estimates for read/write optical disks, which ranged from 10 years to 100 years (Saffady, 1993, Table 5). Today we are much less sanguine: as noted in Figure 3.3, the periods can be signicantly shorter (as little as 3 months) depending on storage conditions and there are other factors to be taken into account. Studies of the life expectancies of optical disks continue, and will need to continue, especially as new types of disk become available and enter into common use (Byers, 2003, p.13). As has already been noted in the section on magnetic tape, there is a clear link between manufacturing quality of optical disks and their longevity. CD deterioration and its connection with manufacturing quality were recognized very early in their existence (Day, 1989). For reasons that include constantly changing production processes and narrow prot margins, the quality of blank CD-R and DVD-R can change from batch to batch and at best . . . can be described as variable (International Association of Sound and Audiovisual Archives Technical Committee, 2004, p.67). When using optical disks in a preservation setting for interim archival use, not for long-term storage it is essential to monitor the quality of new disks before use. This applies equally to blank disks of any reliable brand.
A future for digital artifacts?

The implication of estimates of the longevity of digital artifacts is obvious: the limiting factor for digital preservation is not the lifetime of the media, but the obsolescence of the recording and playback technology, and changing standards. Australian examples illustrate this point. In 1990 the State Library of South Australia produced a videodisk containing about 54 000 images in PAL TV format. The images were played back using a videodisk player and display monitor controlled by a Macintosh computer; the computer hosted a database of image information which was displayed on its own monitor. In 1996 videodisk
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
images were digitized to produce JPEG les with an average le size of approximately 50 kilobytes. The image quality of these was adequate for web display but not for publication or detailed examination. Call number information was added to the images and these modied JPEG images were uploaded to a web server and linked to bibliographic records in the Librarys catalogue. The videodisk continued in service until November 1997. Until recently the JPEG videodisk images were rescanned on demand where customers required highquality digital images. The State Library of South Australia has recently begun a project to systematically rescan the remaining photographs which do not yet have a high-quality master. The images are now scanned as uncompressed TIFF les, at least 3000 pixels along longer dimension and le size at least six megabytes for greyscale and 18 megabytes for colour. The TIFF les are stored on the Librarys 5.6 terabyte RAID mass storage system and on CD-ROMs stored off-site. A backup system for the RAID storage will be implemented, and the CD-ROM backup process may then be discarded. At Geoscience Australia the experience has been that many of the earlier media types used, such as highdensity reel-based magnetic tapes, are unstable, with a life of 10 to 15 years, and the lack of standardization of storage formats for geospatial data has increased the difculty of reading data archived 20 or more years ago. Although Geoscience Australia has developed techniques for recovering data from deteriorated magnetic tapes, the life of tapes is still a limiting factor, and they are forced to refresh their archival data much more frequently than they would like (Tresize, 2002). It is, nevertheless, worth continuing to improve the ability of digital artifacts to store greater quantities of digital data securely for longer periods of time. Although it is well recognized that equipment and software obsolescence, rather than the longevity of digital artifacts, is the limiting factor, it is worthwhile committing research and resources to improving digital artifacts. For example, improvements offer potential reductions in storage costs if storage density is increased, and if the vulnerability of digital storage media to changes in temperature and humidity is reduced. Perhaps more importantly, improvements in media should reduce the frequency with which copying the data (refreshing or migrating it) must be carried out (Hedstrom, 1998, p.197). In addition, the appearance of new kinds of digital storage media and changes, sometimes subtle, in manufacturing techniques or in materials, can affect the length of life of some digital media signicantly (NSF-DELOS Working Group on Digital Archiving and Preservation, 2003, p.10). Ongoing research into digital media is needed. Until we succeed in improving digital media, we need interim responses to the challenges presented by deteriorating digital artifacts. One such response is proposed by Howell, who notes that it is pragmatic to keep crucial pieces of hardware and operating software tucked away from the IT departments upgrading programmes, for a few years at least and provides an example from the State Library of New South Wales, where a 512-inch oppy disk drive was maintained to provide access to legal deposit material in that format (Howell, 2001, p.142). Harvey concluded in 1995 that there are at present too many unknowns to commit digital data to currently-available artefacts for anything other than shortterm storage (Harvey, 1995). This is still the case today. If we choose to preserve digital artifacts, then we do so in the knowledge that the measures we can take to do so are, as yet, short-term expedients.
46
Digital objects more than digital artifacts

At the start of this chapter three sets of challenges for the preservation of digital materials were listed. The rst set, the challenges relating to the nature of media that are used to store digital materials, is described above. This section examines the second and third sets: the challenges arising from the technologies required to create, store and access digital materials; and the challenges to the integrity of digital materials. In the preceding sections the term digital artifacts has been used; this section uses the term digital objects. The distinction between these two terms is helpful in understanding the reasons why digital materials are difcult to preserve. The digital artifact is the physical storage medium, plus the bit-stream recorded on it. The digital object is more than that: it is the bit-stream, plus everything else that is needed to make sense of the bit-stream. For example, a digital object is likely to include information about the format of the data preserved in the bitstream. The distinction noted in Chapter 1, between physical objects and logical objects the terms used by the UNESCO Guidelines (2003) and by Thibodeau (2002) is the key. Physical objects (our digital artifacts) are simply an inscription of signs on a medium (Thibodeau, 2002, p.6) and have no concern with any meaning that is represented by the data in the bit-stream. Logical objects (our digital objects) are processable units . . . according to the logic of some application software. The rules that govern the logical object are independent of how the data are written on a physical medium . . . A logical unit is a unit recognized by some application software. This recognition is typically based on data type [for example, ASCII or other more complex formats] . . . to preserve digital information as logical objects, we have to know the requirements for correct processing of each objects data type and what software can perform correct processing (Thibodeau, 2002, p.7). To preserve the digital object, then, we need to preserve not merely the bitstream, but also the means to process that bit-stream: the access devices that let us read the bit-stream from the digital media on which it is stored; the software that allows us to manipulate and present the information represented by the data carried by the bit-stream; the documentation so that we can understand the data formats used and the software; and the contextual information that is essential to ensure the integrity and authenticity of the information. Ross lists the main factors that can render resources non-interpretable as degradation of the media, loss of functionality of access devices, loss of manipulation capabilities, loss of presentation capabilities, weak links in the documentation chain, and loss of contextual information (Ross, 2000, p.12). His terms are used hereafter in this chapter. As the UNESCO Guidelines for the Preservation of Digital Heritage bluntly put it, Digital materials cannot be said to be preserved if access is lost (UNESCO, 2003, p.22).
Loss of functionality of access devices

The short life-spans of digital media are not the only reason that digital preservation remains a challenge. Another is what is commonly referred to as techno-
Digital objects more than digital artifacts 47
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
logical obsolescence, where the playback devices necessary to retrieve information from the media become obsolete or . . . the software that translates digital information from machine- to human-readable form is no longer available (Hedstrom and Montgomery, 1999, p.1). Even if the digital media on which the bit-stream is stored remains in usable condition and the bit-stream stored on it is intact, the almost certain likelihood is that the drive, software driver or computer will no longer be available to access it. Australia academic Tara Brabazon describes this situation vividly: I still own my rst laptop computer, bought in 1991. It is an Olivetti M316. It functions, although the battery no longer does. It has a 40 MB hard drive, which is not large enough to install a current version of Windows 98, let alone the ability to use the Windows environment to prepare documents. That is probably quite fortunate, as the F key does not work, and most of the letters on the keyboard have been scratched off through excessive use. There is no possibility or space for a modem connection. It does, however, have an expansion slot that is lled with the full-card for my Hewlett Packard at-bed scanner of the same period. It only scans in black and white, and in enormous TIF les, rather than JPEG or GIF. I actually maintained this computer, existing alongside my Sharp PC-M200, until February this year, because I needed the scanning technology and had not yet bought a scanner for my new computer. Once this hardware was bought, the Olivetti computer and scanner became redundant and were taken over by my father, who is teaching himself to use computers through Winders [sic] 3.1 and Word 5. He is managing the technology very well, placing marked stickers over the keys without letters. The F is still causing problems. Between my current computer and the Olivetti, I owned an X-Press 420. The hard drive on this computer had a meltdown in April 1998, and ve of the keys the F (there is a trend here), G, S, L and O did not work. This computer was still living in my house for the years after it was replaced. I would not throw it away, even though it could not be switched on . . . I had a profound hatred for this machine . . . Even though I felt this computer was the digital equivalent of Damien, son of Satan, I could not throw something that had cost me $5570 into the weekly garbage. . . . Burglars robbed my house last December . . . they took the obsolete computer that will not even power up for them (Brabazon, 2000, p.156). Brabazon does not mention another kind of technological obsolescence: the obsolescence of digital storage media, where the media are no longer manufactured and cannot be purchased, drives to access the media are no longer produced, and software needed to control the drives (in the form of device drivers) and to read the data in the formats recorded on the media is no longer written (Rothenberg, 1999a, p.7). The understanding of the general public (as presented by journalists) about digital preservation still has a long way to go. A 2001 newspaper article noted the sale by Booker Prize-winning Australian author, Peter Carey, of his papers and his laptop computer to the State Library of Victoria:
48
The State Library Foundation has bought the laptop as part of a package of Carey papers that effectively tells the inside story of the creation of [the novel] True History of the Kelly Gang . . . The laptop will feature in a Peter Carey section of a permanent exhibition . . . [it] contains at least a dozen drafts of the novel. Umpteen e-mails (Sullivan, 2001). The implication in this article is that keeping the computer itself is sufcient to preserve its contents. We are now much more aware of the naivety of this approach. Different strategies are needed. The information professions have little or no control over this rapid rate of technological obsolescence, which is driven by the constantly changing economic demands of the information technology marketplace. Conway, writing about digitizing, informs us that digital project managers ultimately . . . have no control over the evolution of the imaging marketplace (Conway, 2000). And Rothenberg alludes to another reason, the inuence of those working in the computer industry, who have become inured to the fact that every new generation of software and hardware technology entails the loss of information, as documents are translated between incompatible formats (Rothenberg, 1999a, p.4).
Loss of manipulation and presentation capabilities

The crux of the problem of preserving digital materials is that they are inherently software-dependent (Rothenberg, 1999a, p.8). The bit-stream can represent any of a very wide range of content and formats often text or data, but also images, audio and video, animated graphics, and any other form or format, current or future, single or combined in a hypermedia lattice of pointers whose formats themselves may be arbitrarily complex and idiosyncratic (Rothenberg, 1999a, p.8). This data requires software to interpret it, to turn it into information: This point cannot be overstated: in a very real sense, digital documents exist only by virtue of software that understands how to access and display them; they come into existence only by virtue of running this software (Rothenberg, 1999a, p.8). Meaning is preserved only by preserving the ability to reconstruct streams of bits in a meaningful way that computers and humans can interpret, use, repurpose, and understand at any arbitrary point in the future (Workshop on Research Challenges in Digital Archiving and Long-term Preservation, 2003, p.6). But software is becoming increasingly complex and becomes obsolete rapidly. Besser divides the preservation challenges of this dependence on software into the viewing problem, the scrambling problem and the translation problem (Besser, 2000). Without appropriate software, the digital le cannot be displayed appropriately on the screen, so that all we will be able to see is gibberish. Besser uses the example of les created in WordStar, once common wordprocessing software, which cannot be viewed using todays versions of Microsoft Word. The scrambling problem comprises the complications that result from the use of compression techniques to reduce the size of digital les so that they can be stored and transmitted more economically. These add further complexity to the encoding of les, adding another layer of coding to be recognized and allowed for. The use of lossy data compression techniques discard data, which cannot be recreated. Bessers translation problem refers to the
Digital objects more than digital artifacts 49
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
inability of applications software to read les created for use in its own earlier versions: in fact, he notes, despite the bit-stream being identical, the very reason for converting the le is because we are unable to successfully sustain that applications environment over time (Besser, 2000). Besser also notes what he calls the inter-relation problem. In the digital world material is typically inter-related to other material, a common example being web materials that often contain sections hyperlinked to other web-based materials. This raises an important question: what are the boundaries of a web page? That is, where does one stop preserving? Is it sufcient to preserve only the main web page and not the links to other web sites? (Besser, 2000). Chapter 4 explores this further.
Weak links in the documentation chain and loss of contextual information

Description is the key to converting a digital artifact to a digital object description of three aspects: the bit-stream itself; the systems required to access the bit-stream; and the context in which the bit-stream was created and has been maintained. From a museum curators point of view, objects that have become separated from information about them are worthless, their value being little more than an aesthetic or curio value (Hebditch, 1998, p.viii). Not only must this information about the objects be actively generated and recorded, the systems that exist to maintain this knowledge and to transfer it to the future must also be maintained, for without them the objects become valueless (Hebditch, 1998, p.viii). Museum objects and digital objects have many parallels, not the least of which is the need for documentation about them to remain available so that their meaning continues to be apparent in the future. As the report of the NSF-DELOS Working Group on Digital Archiving and Preservation puts it, There are two problems related to the preservation of digital entities that need generic solutions interpretability and trustworthiness (2003, p.iv). Interpretability is maintained by preserving the means to access the bit-stream and make sense of it, or preserving information about it that allows us to recreate the means. This requires preserving documentation about the data formats used and about the software. An unbroken documentation chain is considered an essential requirement for determining whether a digital object can be trusted. Trustworthiness of the digital object is based on its authenticity and its integrity, that it is what it purports to be, and that it is complete and has not been altered or corrupted (Ross, 2002, p.7). Authenticity and integrity are considered to be core requirements that demand special attention if preserved digital objects are to be trusted. The ease with which digital objects can be altered adds complexity to these requirements, as do some preservation strategies, such as migration and normalization (see Chapter 8) which involve transformations of the original bitstream (NSF-DELOS Working Group on Digital Archiving and Preservation, 2003, p.6). If we cannot be sure of the authenticity and integrity of a digital object, we may not be able to use it effectively. For example, the value of a record of a business transaction in digital form is compromised if it cannot be established that it has not been altered in any way. Figure 3.4 lists the threats to data integrity identied in the UNESCO Guidelines.
50
Natural generation of errors that arise in digital storage systems Breakdown of carriers. Most carrier media have a reasonably short useable life before deteriorating to the point of unreliability for data storage Malicious attack, which may come from system hackers, viruses, staff or outside intruders interacting with the storage system Collateral damage from malicious acts such as terrorist attacks, acts of war or civil unrest affecting buildings or power supplies Inadvertent acts by staff or visitors such as turning off power, throwing out disks or tapes, or reformatting storage devices Natural disasters such as re, ood, or building collapse Business failure.
Figure 3.4 Threats to Data Integrity (From UNESCO, 2003, p.113)
The challenge is to demonstrate the authenticity and integrity of digital objects. This is where the documentation about the object becomes signicant, as authenticity is best protected by . . . documentation that maintains the clear identity of the material (UNESCO, 2003, p.24). For instance, if it can be established that a digital object has always been kept in the custody of an archive in which every change to it has been recorded (an example is changes resulting from migration of the bit-stream the dates at which migration occurred, to which media, and so on), then we can be more secure about its integrity. Jones and Beagrie give further examples: scholars need condence that the references they cite will remain stable, and materials used as evidence in legal situations need to demonstrate authenticity (Jones and Beagrie, 2001, p.25). Ross poses the questions What are the requirements of authenticity and integrity functionality and what can be done to ensure that they are present in digital objects or in the systems that maintain them? He suggests that underpinning authenticity and integrity and their preservation over time are the concepts of xity, stabilisation, trust, and the requirements of custodians and users . . . an authentic digital object is one whose genuineness can be assumed on the basis of one or more of the following: mode, form, state of transmission, and manner of preservation and custody (Ross, 2002, p.7). Many of the concepts that are applied to ensure authenticity and integrity of preserved digital objects come from research and thinking among the recordkeeping community, such as the outcomes of the Functional Requirements for Evidence in Recordkeeping Project at the University of Pittsburgh, and InterPARES (see Chapter 5). Gilliland-Swetland reminds us that the value of an individual record is derived in part from the sequence of records within which it is located and that it can be difcult to understand an individual record without understanding its historical, legal, procedural, and documentary context (Gilliland-Swetland, 2000, pp.16,18). It is also necessary to preserve documentation about other aspects of digital objects in order to ensure their preservation, so that they can be understood in
Conclusion 51
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
the future. Chapter 2 noted the example of recovery of digital data from the Newham Museum Archaeological Service archives, where the lack of documentation meant that much of what was recovered was unusable for archaeological purposes, because the data made no sense without documentation. Gavan McCarthy, Director of the Australian Science and Technology Heritage Centre at the University of Melbourne, provides another example that of databases produced by the State Electricity of Victoria in Australia. When the State Electricity Commission privatized its power generation in the 1990s, its assets were surveyed and it was discovered that about 237 databases or digital information systems had been created. These were analysed to determine which ones had adequate documentation associated with them, or the information within them was intuitive, or implicit enough so that you could understand it and which were, therefore, worthwhile preserving. McCarthy observes that of that 237, I think seven fullled those requirements. Only seven. In many cases we couldnt even gure out what the data was about. There was nobody around who could explain it. (Quotes come from an interview with McCarthy in 2004.) Chapter 6 explores these issues in more detail.
Conclusion
This chapter notes three modes of digital death: instability of storage media, obsolescence of storage and access technologies, and challenges to the integrity of digital materials. The rapid obsolescence of hardware and software is, to a large extent, a result of todays prevailing market-driven ethos, whose highly competitive nature means that product obsolescence is often key to corporate survival in a competitive capitalist democracy (Kuny, 1998). Digital preservation requires that means of addressing rapid obsolescence must be established. (Some of these means are noted in Chapters 6, 7 and 8.) However, rapid obsolescence is not the only threat to the preservation of digital materials imposed by the prevailing market-driven ethos. Another consequence of this ethos is that commercial imperatives seldom coincide with cultural heritage imperatives. Creators of digital materials and other stakeholders may lose interest in their digital output (a business might close down, or a web site might cease to be maintained) with consequences for the future of these materials. Other threats include lack of awareness of stakeholders about digital preservation issues, a shortage of the skill sets needed to preserve digital materials, lack of internationally agreed approaches, a shortage of practical models on which to base preservation practice, and a lack of ongoing funding to address digital preservation issues. These issues are noted further in the rest of the book.
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Chapter 4
Selection for Preservation The Critical Decision
Introduction
Selection in the digital world is not a choice made once and for all near the end of an items life cycle, but rather is an ongoing process intimately connected to the active use of the digital les (Conway, 2000) Based on our knowledge of the past, two things can be said denitively about future library collections: not all recorded information will survive, and we will never be able to predict accurately which information will be in demand by scholars in the future (Smith, 1999, p.2) There are dangers in assuming that current value assessments are a completely reliable guide to future values . . . it is probably better to err on the side of collecting more material rather than less, if the preservation programme can manage it (UNESCO, 2003, pp.7576) Librarians have long acknowledged their responsibility for preserving documents for future use, and have developed criteria and processes for identifying the documents to which they will devote resources to ensure their preservation. The recordkeeping professions, particularly archivists, have also acknowledged this responsibility and have developed a considerable body of theory and practice about appraisal. However, these criteria, processes, theory and practice have been developed and applied to documents in collections that are primarily paper-based collections. They do not automatically translate to digital materials, and need to be revisited and modied to ensure that they can be applied effectively to digital materials. Conway notes the transformation in preservation principles that should be the basis of managing resources in preservation programmes. One of these principles is choice. He notes that
54
Choice is selection . . . Choice involves dening value, recognizing it in something, and then deciding to address its preservation needs in the way most appropriate to that value . . . Selection is perhaps the most difcult of undertakings precisely because it is static and conceived by practitioners as either divorced from present use or completely driven by demand (Conway, 2000). Conway makes one point that is the key to why selection must be reconsidered as we move from old to new preservation paradigms. For digital materials, selection decisions are not a choice made once and for all near the end of an items life cycle, but rather . . . an ongoing process intimately connected to the active use of the digital les (Conway, 2000). The digital mortgage that is the consequence of the selection decisions also needs to be considered: Program costs dont cease when the Web site disappears (Vogt-OConnor, 2000). This chapter considers the important role that selection plays in responsible professional practice of librarians and recordkeepers. It notes selection criteria traditionally applied in library practice and appraisal criteria traditionally used by archives, and then indicates why these selection criteria, developed for physical artifacts, do not translate well when applied to digital materials. It considers what additional factors need to be considered to develop effective selection policies and practices for digital materials, such as the role of intellectual property ownership, the importance of preserving context, and the place of stakeholder input. Emerging frameworks for selecting digital documents are discussed, incorporating some new selection criteria and revised weightings for established criteria. PANDORA, the National Library of Australias web archiving initiative, is examined in Case Study 2 (in the Appendix). This case study illustrates the topic of selection.
Selection for preservation, cultural heritage, and professional practice

Institutions such as libraries, archives and museums have traditionally considered one of their primary roles to be preservation, ensuring that the cultural heritage of the societies they are a part of is available for use by future generations as well as by current users. Preservation is one of the core business activities of these institutions, and they have developed preservation systems and technologies to provide continuing access to cultural heritage materials (NSF-DELOS Working Group on Digital Archiving and Preservation, 2003, p.4). This preservation role has, however, always been one in which major conceptual challenges are embedded. For whom are we preserving? Who exactly are the future users? Because the resources available to cultural heritage institutions (libraries, archives and museums) are seldom sufcient for the preservation of all materials, some selection is required. What are the implications of selecting one item or category of material over another one? How might this selection shape the ways that future generations think about us? Such selection decisions have been the matter of considerable debate, sometimes charged with political and polemical emotion, and they are not easy ones to make. Every choice to preserve is at the expense of something else (UNESCO, 2003, p.73).
Selection for preservation, cultural heritage, and professional practice 55
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Selection decisions what should we preserve?, for how long?, who takes responsibility? are essential in managing collections of heritage materials. They are necessary because there are usually more things more information, more records, more publications, more data than we have the means to keep (UNESCO, 2003, p.73). The networked environment has exacerbated this situation. In the pre-digital print environment the business of publishing provided (and continues to provide) some ltering, some quality assurance of the product, through mechanisms such as publishers readers who make decisions based on quality, relevance to a dened public, and saleability. As noted in Chapter 1, increasing quantities of information are being produced digitally, and it is easier for individuals and organizations to make available to the public to publish without any intervening quality assurance measures. The ease of mounting information on web sites has meant signicant increase in the amount of information readily available, but it has also meant signicant variability of its quality. The inevitable result has been an even stronger requirement to select materials for preservation purposes in the face of insufcient budgets, expertise and facilities. Digital preservation will be expensive, the Cedars Project reminds us in response to the question why do we need to select? Of one fact we can be sure: digital preservation will require more resources than preserving print material (Russell, 1999). And, unlike non-digital material such as paperbased artifacts, where there is a comfort zone (Jones and Beagrie, 2001, p.30) in which to make selection decisions before the materials deteriorate, the time frame in which decisions about whether or not to preserve digital materials is very short. Selection decisions for cultural heritage materials and this increasingly includes digital materials have been the focus of considerable debate. The issues have been articulated most clearly by the recordkeeping professions. Cook summarizes many of them. Archives, he suggests, are: a source of memory about the past, about history, heritage, and culture, about personal roots and family connections, about who we are as human beings and about glimpses into our common humanity through recorded information in all media, much more than they are about narrow accountabilities or administrative continuity (Cook, 2000, p.5). But there are many dangers: Memory is notoriously selective in individuals, in societies, and, yes, in archives. With memory comes forgetting. With memory comes the inevitable privileging of certain records and records creators, and the marginalizing or silencing of others (Cook, 2000, p.5). Views such as the ones Cook articulates are not the only response to selection emanating from the recordkeeping profession. A more traditional view is based on a clear distinction between the activities of records managers and the work of archivists. Records managers select records for retention based on risk avoidance, market opportunities, or desires to avoid embarrassment or accountability but this approach inevitably will privilege the needs of business or government in terms of the issues that get addressed, the allocation of resources, and the long-term survival of records (Cook, 2000, p.8). The records that survive into
56
the future will reect the concerns of administrators, rather than the full range of human experience. Recordkeepers, suggests Cook, need as a profession to remind [themselves] continually of the fate of records left to White House presidents and Soviet commissars, South African apartheid police forces and Canadian peacekeepers in Somalia, rogue Queensland politicians and the American Internal Revenue Service (Cook, 2000, p.8). However, in the need to select, the selectors must inevitably bring to their decisions their own beliefs and values. They move from a strictly objective custodial role to becoming themselves creators of social memory through the active formation of the archival heritage (Cook, 2000, p.5). Responsible selection practice in preservation must aim to minimize bias in the value judgments that are inevitably contained in the decisions about what to select for preservation. Selection disenfranchises some groups, as our experiences with non-digital collections have shown; an example is that history has moved from being that of signicant individuals (usually men) to a wider view based on the records of other groups such as women, the poor, and indigenous cultures. Brabazon has argued that digitization is actually and actively reinforcing the social exclusions of the analogue world because of the digital divide between the information rich, information poor (Brabazon, 2000, p.154) and that even more of the memory of socially excluded groups will dissipate in the ephemeral winds of cyberspace (Brabazon, 2000, p.155). Should we save everything? Current thinking is that we cannot, because we lack sufcient resources. But there are other reasons why the answer to this question is no. Burrows provides the example of academic libraries, whose collections are, he maintains, a product of the rational and scientic master narrative of the Enlightenment. This has been challenged by the new humanities and social sciences which increasingly demand that libraries collect more of what used to be dismissed as ephemeral and popular. Added to this is our pervasive faith in the power of technology to solve every problem which leads us to believe that everything could and should be preserved (Burrows, 2000, p.146). Whether or not this view is realistic and technologically feasible, the current state of development of digital preservation is that we still have to pose the question of what really matters. Burrows points to our current predicament: We face a real danger of being caught between two contradictory imperatives: the universalist, Alexandrian demands of cultural relativism, and the limitations imposed by technology, cost and organizational structures. Cultural relativism and postmodernism make it increasingly difcult to reach any consensus on dening what really matter, while insufcient funding and inadequate technology make it very difcult to achieve the goal of archiving everything (Burrows, 2000, p.148). He suggests four selection scenarios. The rst is picking the low-hanging fruit, that is preserving what is easiest to preserve (this metaphor is used in UNESCO, 2003, p.79). This may be unsatisfactory because what is easiest to preserve is unlikely to be what is of particular value. The second is to let the marketplace
Selection criteria traditionally used by libraries 57
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
decide. In this approach, the denition of value is likely to be linked to commercial considerations rather than to long-term societal value. The third is to base selection decisions on the type of material, but Burrows suggests that this scenario raises more questions than it answers: are websites more important to preserve than Usenet postings? The fourth is to select on the basis of the most endangered material (Burrows, 2000, p.148). None of these are, however, completely satisfactory as a basis for making realistic decisions that are professionally responsible, although some of them have been applied to digital materials on a small-scale basis, so that we can develop expertise and scalable strategies and practices for digital preservation. Where might we look for guidance on these issues? This chapter suggests that selection criteria based on library practice and applied to non-digital materials are insufcient, and that more appropriate guidance is to be found in appraisal theory and practice derived from the recordkeeping discipline. Archivists, notes Burrows, have already done considerable work in rethinking selection criteria for the retention of records in a digital environment (Burrows, 2000, p.149). However, all matrices of selection criteria, regardless of their disciplinary origins, share common problems. One problem is the difculty of determining value. Current assessments of value are not always a useful indicator of future value: the UNESCO Guidelines for the Preservation of Digital Heritage (UNESCO, 2003, pp.7576) provide the example of remote sensing data from earlier decades, which is now being used to assess environmental damage. Another problem is determining how much to preserve. There is a growing perception that we need to keep more, rather than less, of the digital environment, partly because we have the technology to do so (such as that used by the Internet Archive) and partly because our abilities to interrogate and get new meanings from digital data are increasing as new tools such as data visualization and data mining software are developed (Kerry, 2001, p.11, quoting Ross, 2000, p.12). There is also an increasing perception that generalized selection criteria are not appropriate, and that sectoral differences should be further investigated and selection decisions made on this basis (Kerry, 2001, p.16).
Selection criteria traditionally used by libraries

Libraries, with their long history of preserving documentary heritage, are still more attuned to making preservation decisions about the artifacts (the physical objects) in their collections than about the less tangible digital materials. Many of the policies developed for the selection of artifacts also apply to digital materials. Certainly the key issues remain the same: on what basis do we select? What are the implications for current and future users? Who might the users of the future be? Who decides how much should be kept? Who keeps it? (Much of the rest of this chapter is based on Harvey, 2003). The most important principle that guides most library selection practice is current user need. Many of the policies developed for selection by libraries are aimed at meeting the needs of their user communities. (National libraries are a signicant exception to this in their use of legal deposit mechanisms to acquire comprehensive national documentary resources collections.) Secondary principles, such as decisions to develop particular subject areas where user needs are of lower priority, may modify these policies. However, selection guidelines
58
based on meeting current user needs do not automatically apply to selecting material for preserving in the future. Criteria that have been developed by libraries to select artifacts for preservation vary little. The characteristics of artifacts which these criteria address are age, evidential value, aesthetic value, scarcity, associational value, market value, and exhibition value (Task Force on the Artifact in Library Collections, 2001, pp.9,11). There are some other perspectives on these criteria; for instance, the fact that an artifact is an original and not a copy may be strongly associated with its evidential value and its exhibition value. Selection criteria are sometimes described in other terms. For example, the criterion of evidential value is variously described as manufacturing or evidential value (Pacey, 1991, p.189), rarity in terms of informational content (Lyall, 1993, pp.7274) and scientic or research signicance ((Signicance), 2001, p.11). More examples are given in Harvey, 2003, Table 1. Occasionally additional criteria are used: examples include fragility or condition (Pacey, 1991, p.189) and whether the artifact is held in community esteem ((Signicance) 2001, p.11). In recent years the practicalities of preservation are appearing more frequently, among them the management plan (is there a plan which reects the signicance of the documentary heritage? (Edmondson, 2002, pp.2223)) and assessment of physical condition (is its survival in danger? (Edmondson, 2002, pp.2223)); is this item in danger of being lost? (Harris, 2000, pp.206224)). Constraints to preservation resulting from intellectual property issues have also made an appearance, as in Harriss are there any constraints caused by copyright laws? (Harris, 2000, pp.206224). To summarize, traditional library selection practices for preservation are based on the preservation of items (or artifacts) in their original formats, according to ve key criteria: evidential value, aesthetic value, market value, associational value, and exhibition value. Additional criteria are often also used, although there is less consistency in their application: physical condition, resources available, use, and an ill-dened, but nonetheless well-represented, category of social signicance (is the item held in community esteem?). Although this summary sounds straightforward and its criteria clear cut, in practice many problems arise with their interpretation, most of which relate to the difculty of determining what value is. What were the grounds for detecting in favor of one object and against another? How can libraries cope with the fact that the value of the artifact is never quite the same to different researchers? Artifacts are cultural variables that are viewed and used in a given culture at a particular moment (Task Force on the Artifact in Library Collections, 2001, p.12). Ayres asks a question about Australian literature to illustrate this point: to ask how well Australian archives represent Australian literature is to ask questions about what Australian literature is. Is there an Australian literary canon? Interrogations of canonicity have raised important questions about high versus popular literature, gender and cultural biases, aesthetic versus political concerns, notions of Australianness, and literature as cultural and economic production (Ayres, 2001, p.33). Libraries havent always got this right in fact it is impossible to get it right, given the context of changing perceptions of value and the uid dynamics of intellectual inquiry (Task Force on the Artifact in Library Collections, 2001, p.5). As an example, consider the changing status as research materials of Australian romance novels, a genre which has been banish[ed] . . . from the cultural record of the nation (Flesch, 1996, p.190;
Appraisal criteria traditionally used by archives 59
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
2004). Other problems arise when determining if and when a surrogate is just as suitable as the original artifact. Copies of artifacts lose some of their information each time they are copied a microlm can never capture some of the information that resides in the physical structure of a book. This is known as generational loss. Libraries have tested many approaches to selection for preservation. For some of these the aim has been to reduce the amount of intellectual input into the selection process, or to expend that effort only once so that others do not have to duplicate it. The great collections approach was based on selecting artifacts whose content was valuable intellectually and which were physically fragile, but this is labour-intensive. Similar approaches involved the determination by a bibliographer or panel of experts of the core literature in a eld, but this, too, is labour-intensive and requires a high level of bibliographic control in the eld. User-driven selection occurs when an item requested by a user is in poor physical condition and is treated when attention is called to it (Cox, 2002, pp.9798). Smith (1999, pp.911) provides more details of these approaches. None of these approaches is considered to be entirely appropriate to the selection of artifacts for preservation. The same reservations apply to their application in the digital environment.
Appraisal criteria traditionally used by archives

The archival community has well-developed theory and a set of practices for selection, known as appraisal. This is well worth investigating for its suitability in a digital context. Appraisal practice is signicantly more developed than library selection practice, so it is feasible that the deeper theoretical basis of appraisal can inform our thinking about what digital materials are worth preserving. Appraisal lies at the core of archival practice. It is traditionally dened as the process of evaluating records to determine which are to be retained as archives, which are to be kept for specied periods and which are to be destroyed (Ellis, 1993, p.461). Traditional appraisal of archival materials applies criteria similar to those used in library selection. Archival value encompasses several narrower values: administrative (usefulness for the conduct of business), scal (usefulness for nancial business), legal (worth for conduct of legal business), intrinsic (the inherent nature of the material, its signicance as artefact), evidential (its value as evidence of the record creators origins, functions and activities), and informational (usefulness for more general research purposes because of the records information content) (Tibbo, 2003, pp.2930). What most strongly distinguishes archival appraisal practice from library selection practice is a stronger recognition of, and emphasis on, the importance of context. Cox suggests that from an archival perspective, context is crucial to understanding information or evidence (in any form) (Cox, 2002, p.53). The emphasis placed on context is the key difference between library practice and archival practice in selection. The value of archival theory and practice for developing effective practice in the digital environment is noted by Gilliland-Swetland:
60
The archival perspective brings an evidence-based approach to the management of recorded knowledge. It is fundamentally concerned with the organizational and personal processes and contexts through which records and knowledge are created as well as the ways in which records individually and collectively reect those processes. This perspective distinguishes the archival community from other communities of information professionals that manage decontextualized information (Gilliland-Swetland, 2000, p.2). This evidence-based approach with its emphasis on context is likely to be more helpful in developing effective selection criteria for digital materials than is selection practice based on the traditional library criteria. Digital materials can be more effectively handled by considering them, at least in part, as records. For example, like records, digital materials need to have preservation decisions made about them at a point closer to their creation, whereas in library practice decisions of this kind can be made many decades after the items production. Library-based practice has already concluded that documentation and metadata for digital materials, both of which provide contextual information, are requirements for their effective preservation. Closer investigation of appraisal concepts is likely to provide further guidance. The Archives Association of Ontarios brief statement about appraisal (Archives Association of Ontario, 2001) poses six questions: 1. What is the administrative, evidential or informational value of the records to the organization? 2. Do the records meet the terms of your mandate and acquisition policy? 3. Are the records primary or unique? 4. Is the information in the records duplicated in another set of records? 5. Can the records be properly preserved? 6. Can the records be made available? Two of these (questions 5 and 6) are concerned with practicalities: are the resources (technical, nancial, and human) available to preserve the records? Are access restrictions so rigid that the records cannot be made available to users in a realistic and timely way? The rst question is at the core of appraisal. By linking the administrative, evidential and informational value of the records to an organization, it clearly suggests the context aspect: the context the organization is dened. Cox further indicates that a record is a specic entity. It is transaction oriented. It is evidence of activity (transaction) and that evidence can only be preserved if we maintain content, structure, and context (Cox, 2001, p.6). How might we interpret this for application to materials that are not records of transactions, such as the kinds of materials, including digital materials, usually handled by libraries? Cox assists us: Structure is the record form. Context is the linkage of one record to other records. Content is the information, but content without structure and context cannot be information that is reliable (Cox, 2001, p.6). We need to identify the equivalents of content, structure, and context for digital materials that are not records. However, the reality is not as straightforward as this. Appraisal, notes Piggott, is not free of bias and subjectivity; its results reect the cultural and other values of the time. He describes the great trinity mystery:
Why traditional selection criteria do not apply to digital materials 61
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
apart from some exceptional cases, it is beyond our resources and power to keep all records; which is a pity, because beyond their original use, all records conceivable have their uses; weve come to expect unexpected uses and yet it is almost impossible to accurately predict future use, and when we try, the passage of time can cause serious havoc with appraisal judgments (Piggott, 2001). Appraisal theory and practice are themselves being continually reappraised. For example, the development of the macro-appraisal concept is one attempt to rene decisions made about value. It is nothing more than deciding (what records to create and) how long they should be kept or deciding what archival records to collect, by rst mapping the territory; by rst identifying and analyzing the theoretical documentary universe . . . Only after that intensive research based examination of societal or business functions . . . should you do micro-appraisal by identifying the key records (Piggott, 2001). This approach, it is suggested, helps cope with greater quantities of records. All of the traditional appraisal values now come into play plus the secondary factors such as existence of control records, costs, and preservation (Piggott, 2001). The denition of appraisal is being reconsidered to accommodate digital records, so that appraisal decisions are made at or near the time that records are created. The ISO standard Records Management Guidelines, which is based on an earlier Australian Standard for Records Management, does not dene appraisal but notes that capture is the the process of determining that a record should be made and kept . . . It involves deciding which documents are captured, which in turn implies decisions about who may have access to those documents and generally how long they are to be retained (AS ISO 15489.2, 2002, 4.3.2 the italics have been added).
Why traditional selection criteria do not apply to digital materials

Some argue that there is no difference between selection for preservation of digital materials and selection of non-digital materials, that the selection of digital heritage is conceptually the same as selection for non-digital materials: the same principles and practices, based on existing policies and procedures, apply, although they may need some adjustment (UNESCO, 2003, p.73). This may be too simplistic a view. It could be argued that the different emphases required are so great that the concepts are altered signicantly. For instance, the increased emphasis we must place on whether we have the technical capability to preserve digital materials means that the practicalities become more important, whereas in the past the information content of most materials in the collections of libraries and archives would not be altered if no actions were taken. What are the differences that affect selection decisions? As noted in previous chapters, digital materials are more vulnerable than material in traditional
62
formats, such as paper. One of the major differences is the cost of maintaining these more vulnerable materials. While major resourcing issues are attached to any decisions we make about which materials we will retain, the issues are greater for digital materials (the digital mortgage noted earlier in this chapter). For example, the cost of preparing a digital object for preservation could be signicantly higher than the cost of preparing a book for addition to a librarys collection, and (assuming a xed budget) fewer objects can therefore be processed (Weinberger, 2000, p.65). Another difference is that preservation strategies and techniques developed in a pre-digital era do not always apply in the digital environment: in Cloonans words, digital objects force us to preserve them on their own terms (Cloonan, 2001, p.237). For example, selection decisions made about digital materials probably need to take into account whether or not the equipment to read them is available; this is not typically the case for non-digital materials. We need to develop appropriate preservation strategies and techniques for the digital environment. A third difference is the requirement that conscious preservation decisions need to be made about digital materials early in their existence. Burrows compares the pre-digital environment with today: In the pre-digital era, selection necessarily preceded preservation. Once a book or journal had been acquired, a decision could be made about its long-term retention. The mere act of placing it in a library was often seen as a sufcient method of preservation in itself. In the digital era things are very different. Digital records dont just survive by accident, observes Margaret Hedstrom. As a result, a lack of preservation is tantamount to de-selection (Burrows, 2000, p.152). As the UNESCO Guidelines succinctly state, it may not be possible to wait for evidence of enduring value to emerge before making selection decisions (UNESCO, 2003, p.74). More differences between digital and non-digital materials that need to be considered when developing selection criteria are listed in the UNESCO Guidelines (UNESCO, 2003, p.74). For digital materials the quantities to be assessed may be signicantly greater, and their quality may be more variable as, unlike printbased publications, they do not always go through a quality assurance process imposed by their publisher. We also face the challenges posed by new genres being developed, especially those that incorporate linked resources. There is too the question of exactly which attributes of digital materials should be preserved (noted in more detail in Chapter 5. Yet another issue is the difculty of determining ownership of intellectual property rights for some digital materials. Any effective process and criteria for selection of digital materials for preservation must take account of these differences.
IPR, context, stakeholders, and the continuum approach

Viable selection criteria for digital materials need to take account of many factors that differentiate them from non-digital materials. Intellectual property rights
IPR, context, stakeholders, and the continuum approach 63
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
assume greater importance, legal deposit for digital materials is slowly becoming a reality, and there is a need to preserve much more contextual information than is thought necessary for most non-digital materials. Stakeholders potentially play a much greater role. A continuum approach, borrowed from recordkeeping theory, shows promise in assisting us to develop effective selection principles.
Intellectual property rights and legal deposit

There are already well-established legal frameworks for preservation of traditional (usually non-digital) material that enters the collections of libraries. These include provisions in copyright law and legal deposit legislation. For instance, copyright laws are likely to include a provision that an item can be copied without specic approval from the copyright owner if the copies made are for preservation purposes. No similar blanket provisions apply to digital materials. For them the legal framework is in a state of ux as it attempts to accommodate new information society paradigms and practices (Jones and Beagrie, 2001, p.32). Copying digital material for preservation purposes the very basis of the digital preservation strategies of refreshing, migration, and emulation can infringe intellectual property rights as they currently stand for some material (Muir, 2004a, pp.7677). Some countries have recognized this in their copyright legislation, as in the United States Digital Millennium Copyright Act (1998) and in a 1997 amendment to the Canadian Copyright Act, which allow digital materials to be copied if their format has become obsolete (Muir, 2004b, p.72). For Australia, the Copyright Amendment (Digital Agenda) Act 2000, which came into effect in March 2001, allows digital copying of material held in libraries and archives for preservation purposes. Some countries have data protection acts that may also present barriers to preservation (Jones and Beagrie, 2001, p.34). Commenting on trends in intellectual property rights, Beagrie notes that, as the economic value of intellectual property increases and investment in it expands, legislation becomes more restrictive: for example, copyright protection periods have been extended. He comments that the needs of memory institutions for legal exceptions to undertake archiving are often overlooked or not sufciently understood and are overridden by commercial imperatives to protect intellectual property (Beagrie, 2003, p.3). There are at present no simple solutions to this issue. Strategies include establishing a dialogue with intellectual property rights owners so that they are fully informed about the issues and about what is required to allow the preservation of digital materials (Jones and Beagrie, 2001, p.32). Intellectual property rights issues may inuence selection of digital materials for preservation. If the rights are so restrictive that there is no real possibility of access to the material being made available in the future, then it would probably be pointless to expend resources on its preservation (UNESCO, 2003, p.77).
Legal deposit legislation

One of the powerful tools that some libraries are able to use in the preservation of traditional documentary heritage materials is legal deposit legislative provisions, through which material comes automatically to a designated library, without the expenditure of considerable effort and resources to acquire that material. Digital materials are not covered by legal deposit legislation, which
64
predates the digital era, except in a handful of countries. Legal deposit is conceptually antithetical to selection: it implies that no selection is made, but that all materials dened by the legislation are automatically provided to the deposit library. Legal deposit is noted here, however, because it addresses one of the issues faced in preserving digital materials: getting hold of them in the rst place. Legal deposit legislation for digital materials is often considered as an important step in a viable national preservation programme for digital materials. Without it, preservation of digital materials is necessarily ad-hoc (Kerry, 2001, p.8). Countries that have enacted legislation covering all forms of digital publication include Denmark (1998), New Zealand (2003), and the United Kingdom (2003). Other countries have legal deposit legislation that covers some digital materials, usually static publications such as those issued on CD-ROM. Without legal deposit legislation, there is some potential in negotiating less comprehensive arrangements such as the National Library of Australias Voluntary Deposit Scheme for Physical Format Electronic Publications (NLA, 1999?) and its Use of Australian CD-ROMs and Other Electronic Materials Acquired by Deposit (NLA, 2001). But these are more time-consuming and resource-intensive arrangements and the National Library of Australia is keen to see legal deposit legislation for digital materials enacted in Australia so that it can carry out more effectively its role to preserve Australias documentary heritage. Such legislation would enable the National Library of Australia to collect important online publications without negotiating with the publisher, and consequently more national resources of signicance would be preserved and available for use in the future, and more cost-efcient and effective work procedures would be possible for both publishers and the Library (Gatenby, 2002b). Australias neighbour, New Zealand, has enacted legislation in 2003 which provides for the deposit of physical format documents and for the copying of Internet documents and the National Library of New Zealand is developing its procedures to handle the resulting acquisition of digital materials (Legal deposit, 2003).
Context and community

The importance of preserving contextual information has already been noted in earlier chapters. It is not, as noted in Chapter 1, simply the bit-stream that we wish to preserve, but also the additional information and tools needed to access and understand that bit-stream. Selection processes must take this into account. If the contextual information is not available, then the digital materials are unlikely to be selected. The UNESCO Guidelines suggest some specic instances: Where digital materials can only be understood by reference to a set of rules such as a record keeping system, database or data generation system, or other contextual information, selection processes must identify the documentation that will also need to be preserved (UNESCO, 2003, p.76). Preserving contextual information, challenging in itself, becomes more so when the aims of preservation programmes are also taken into account. Every programme is itself working within its own context. For national libraries that
IPR, context, stakeholders, and the continuum approach 65
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
context is the nation; for a business archive, the company that it is established to serve; for a university, its faculty and students. The community that the programme serves is likely to impose its own requirements. This directly affects the quantity and nature of digital materials that are selected for preservation: in the case of a national library it is a wide range, whereas for a universitybased programme it will be narrower, perhaps restricted only to the intellectual output of its faculty and research students. The necessary inuence of context on selection for preservation is exemplied in this recommendation of an Archaeological Data Service (ADS) study into the use of archaeological data in digital formats: strategies are needed to identify which datasets should receive the highest priority for preservation. Such strategies should be determined by the information needs of archaeologists and must be able to react to those projects that are most in danger of being lost (Condron et al, 1999, section 11.7). As well as the nature and extent of material that is preserved, the community will also dene what kind of contextual information is also collected and preserved. For some communities it may be sufcient to see only a passive rendition of the digital materials: a screen shot or other visual presentation; perhaps a PDF version is all that is required. For others, it will be necessary to retain sufcient contextual information to allow the digital materials to be searched or manipulated. Users need the option of interrogating old data to produce new results . . . Some programmes may even have to ensure users can run old simulations, play old computer games, or view digital art in ways that reproduce the original experience rather than a speeded up experience that later technologies may provide (UNESCO, 2003, pp.7778). For other communities, there must be enough contextual information of the right kind to demonstrate that the authenticity of the digital materials has not been compromised. Selection criteria for determining digital materials to be preserved need to take account of this contextual information. Chapter 5 examines further the questions surrounding what attributes need to be preserved.
Stakeholder input
Chapter 2 noted that stakeholders would play a greater role in digital preservation than they played in the past in the preservation of non-digital materials. An example is the need to communicate with intellectual property rights owners so that they understand the preservation implications of the control they wish to exert over digital materials. The UNESCO Guidelines suggest that because producers of digital materials are well placed to understand why digital objects were brought into being, their essential message, and the relationships between objects and their context they are likely to play an important role in selection decisions. If that information is not captured from the producer it may be too difcult to reconstruct it later (UNESCO, 2003, p.76). Another example is the role of the individual in preserving popular culture. Where previously institutional collectors (libraries and archives) were unlikely to collect the more ephemeral materials that constitute popular culture, there were
66
myriad alternative sites where ephemeral material was stored, such as the family home [where we might expect to nd] a light sabre, toy dalek, Duran Duran posters. [But] the digital era reduces the number of fanbased archivists. Subsequently forfeited is the spectrum of interests and ideologies that construct the popular memory of a culture (Brabazon, 2000, p.157). Increasing engagement of some community sectors in the selection of digital materials for preservation is also envisaged. A summary of the key areas for research in digital preservation in Australia includes the statement that more work needs to be undertaken on a sectoral basis on establishing signicance assessment policies . . . with clearly established priorities (Kerry, 2001, p.16). Greater stakeholder input is being encouraged in many areas. For example, a stakeholder consultation approach to appraisal is being introduced by the National Archives of Australia (Schwirtlich, 2002, pp.6061). Peer review has been used to determine the value of retaining scientic data sets (National Research Council, 1995, p.34). A related idea is authority by community, a concept used by Janes in relation to the authoritativeness of web sites. Janes provides the example of the Internet Movie Database (www.imdb.com) which requests that errors and omissions be reported to the IMDB database managers, who will check and incorporate if relevant. Janes notes that weve grown used to the notion that a sites content could be dynamic; now it may well be that its authority is as well (Janes, 2003, p.92). Another example is the Wiki (dened in perhaps its best-known manifestation, the Wikipedia a free-content encyclopedia that anyone can edit as a website (or other hypertext documents collection) allowing users to add content, as on an internet forum, but also allows anyone to edit the content (en.wikipedia.org/wiki/Main_Page). This suggests greater input into the selection process by users and other stakeholders. Members of communities have a stake in keeping digital les accessible. In universities, faculty members play an expanding role by participating in the establishment of their institutions e-repositories (Smith, 2003, pp.1314). Discipline-based approaches, also present in the pre-digital preservation paradigm, are proving effective, as shown by examples such as the Astrophysics Data System (adswww.harvard.edu) and Bright SPARCS (www.asap.unimelb. edu.au/bsparcs/bsparcshome.htm). For scientic data, there appear to be significant differences in how appraisal criteria need to be applied in different disciplines, a contention that needs further investigation (Hodge and Frangakis, 2004, p.59).
Value of the continuum approach

The point has been made in this book several times already that selection decisions about which digital materials to preserve need to be made at an early stage in their existence. The continuum approach, developed for application in recordkeeping, shows promise in assisting us to develop effective selection principles. The continuum approach (explained in more detail in Upward, 2005) is essentially a way of thinking about the life of a record, from its creation onwards. It was conceived to get around the issues associated with the traditional split between records and archives. The recordkeeping community has been at the forefront in developing responses to digital preservation issues. For electronic
Developing selection frameworks for preserving digital materials 67
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
records, recordkeepers can no longer wait passively at the end of the life cycle for records to arrive at the archives when their creators no longer wanted them or were dead (Cook, 2002, p.2). Signicant records must be identied early in their life so that management and preservation decisions that will ensure ongoing access to the critical aspects of those records, such as the information content and the attributes that determine their authenticity, are made right from the start. This, suggests Weinberger, will ease the selection of digital objects for preservation (Weinberger, 2000, p.67). JISC (the UK Joint Information Systems Committee) has recognized the value of a continuum approach. It is supporting a life-cycle approach to the management of digital resources (Beagrie, 2004).
Developing selection frameworks for preserving digital materials

The point that traditional library preservation paradigms do not transfer well to the digital environment has been made already. The critical reasons why they do not stem from the fact that issues such as media instability and technological obsolescence shorten the timeframe within which decisions about the future of digital materials must be made. The physical condition of artifacts drives many traditional preservation decisions (Gertz, 2000, p.98) but this factor is of little or no importance when applied to digital materials. In the past we have decided what to preserve by keeping everything that we can and waiting for the signicant information to rise to the surface as time, people and events dictate its importance (Howell, 2000, p.129). This approach has also been applied in the digital environment, the most notable example being the Internet Archive. There remains doubt about whether this approach is sustainable as the quantity of digital materials increases. The persistence of this approach avoids what is very likely to be the real issue: how to determine signicance. Most of the literature published to date about selecting digital materials is about the selection of analogue material for digitization, not about the selection of digital material for preservation. These are not the same. The principal reason for selecting material to digitize is usually to improve access to that material, which may still be subject to pre-digital preservation regimes, whereas selection for preservation is aimed at long-term retention of access to information that is already, and only, in digital form. Digitizing programmes are usually access-driven, with preservation as a by-product (Gertz, 1998). Despite those differences, criteria for selecting materials to digitize can provide useful advice for developing selection criteria for preservation. Many of the traditional selection criteria apply, although the emphasis is different: for example, greater prominence is given to user demand and intellectual property rights (Gertz, 2000, pp.9899). Gertz has identied the criteria most frequently cited as: Does the item or collection have sufcient value to and demand from a current audience to justify digitization? Do we have the legal right to create a digital version? Do we have the legal right to disseminate it? Can the materials be digitized successfully? Do we have the infrastructure to carry out a digital project?
68
Does or can digitization add something beyond simply creating a copy? Is the cost appropriate? (Gertz, 2000, p.104). The Cedars Project Team report includes one of two especially helpful statements about selecting digital materials for preservation (Cedars Project Team, 2002). It denes a digital objects Signicant Properties as the level of content and functionality retained. These signicant properties are not empirical; they require judgments to be made by organizations about how these signicant properties apply to their user communities and to the organizations preservation responsibilities. The Cedars Project Team report provides an example: for a text in PDF, the decision is made that the text, not the format, is signicant, so information about the PDF format does not need to be stored (Cedars Project Team, 2002, p.15). The report places high importance on negotiating intellectual property rights before any other preservation actions occur. It suggests that selection decisions for digital preservation must be pragmatic and based on the estimated value of the material, the cost of storage and support mechanisms, and the production of metadata to support the material (Cedars Project Team, 2002, p.53). Primary criteria are proposed: currently high use the type of material (typically commercially published scholarly works) that we would expect to preserve if it were published in traditional printed format tied to the long-term or cultural interests of the organisation (Cedars Project Team, 2002, section 4.5.1). A typology of digital resources proposes additional criteria (Cedars Project Team, 2002, pp.109110) including: Legal and IP issues: (for example, legal status IP rights need to be negotiated for preservation purposes) Format issues (for example, some formats harder to preserve than others) Technical issues (for example, Technical capacity to preserve may be lacking in a library; some technical environments may be easier to preserve than others). The Decision Tree for Selection of Digital Materials for Long-term Retention, in the Digital Preservation Coalitions handbook on the preservation of digital materials (Jones and Beagrie, 2001, section 4), also assists. This decision tree rst poses questions relating to selection policy: is there an institutional selection policy? Does the material t into it? Is the material of long-term value? A second group of questions is about legal and intellectual property issues: have acceptable rights been negotiated? Can they be? Technical questions form a third group of questions: can you handle the le format, now and in the future? Can the material be transferred to a more manageable format? The existence of documentation and metadata form a fourth group: has sufcient been supplied? To summarize: changes are required to the selection procedures developed for non-digital artifacts to accommodate digital materials. Some of the traditional factors assume greater importance, but new factors must also be considered for digital materials. The digital selection frameworks developed to date still place high priority on criteria for determining value, but do strongly emphasise other
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
criteria: the legal and intellectual property rights governing a resource, whether we have the technical ability to preserve it, the costs involved in preserving it, and the presence of appropriate documentation and metadata. Figure 4.1 summarizes these changes.
Towards a new decision-making framework

Weinberger poses three questions that need to be answered when selecting digital materials for long-term preservation: Do we want to preserve this digital object? May we (do we have the legal rights to) preserve this digital object? Can we (do we have the technical ability to) preserve this digital object? (Weinberger, 1999).
Traditional Selection Criteria Value Evidential Aesthetic Market Associational Exhibition Informational Threat to object Fragility Management plan? Heavy use Held in community esteem? Copyright
Criteria Applied to Selection for Digitizing Sufcient value to current audience Does digitization add value?
CEDARS/DPC Decision Tree Signicant long-term value?
Physical condition Resources available Use Social signicance Legal rights Format issues
Is infrastructure available? Is cost appropriate? Current demand Currently high use Tied to long-term interests of organization Rights to digitize Rights to disseminate Can it be digitized successfully? Legal status IP rights Type of material (can it be digitized successfully?) Technical ability to preserve? Can le format be handled? Selection policy? Sufcient available?
Technical issues
Policies Documentation Figure 4.1 Selection Criteria Categorized
70
We need to expand upon this starting point. As already noted, increasingly nontraditional criteria, such as the legal and intellectual property rights governing a resource, whether we have the technical ability to preserve it, the costs involved in preserving it, and the presence of appropriate documentation and metadata, are becoming central to decision-making in selection of digital materials for preservation. Many of these factors have been recognized for some time by the audiovisual archiving community and have been incorporated into their selection procedures. The process for determining what to migrate to digital format at the television archives of a German public service broadcaster, the Sdwestrundfunk, Stuttgart, is typical. This organization operated a cautious policy on de-selection because their experience indicated that future requirements were not easy to predict. Nonetheless, it would be too expensive to transfer everything to digital format, so a more sophisticated prioritization schedule was developed. In addition to criteria based on content and form (documentary, magazine, spot and so on) the archival worthiness of the material is determined, based on the cataloguers assessment. Also considered are broadcast rights (where these exist only for a limited time, the material is not selected), physical carrier condition (good; at risk because of brand, age, frequent use; showing signs of deterioration) and status of the copy (is it a copy, a copy where the original still exists, etc). This process is time-consuming, but nevertheless it is to be strongly recommended (Lacken, 2001, p.25). In Phase 1 of its investigations the InterPARES project addressed the question How do we select electronic records for preservation?. It determined that records in digital format should be selected for long-term preservation on the basis of continuing value, authenticity, and feasibility of preservation. The authenticity of records should be established using the InterPARES Benchmark Requirements for Assessing the Authenticity of Electronic Records. The feasibility of preservation should take into account current and anticipated capabilities and projected costs. Appraisal of records in digital form should be carried out as early as possible, and ideally should be built into the design of record-keeping systems. Assessment of authenticity and feasibility may be carried out later. Appraisal decisions should be monitored regularly to ensure that the information kept about the records is valid: the recommendation is to ensure that changes to records and their context have not negatively affected their identity or integrity nor the ability to preserve them (US-InterPARES Project, 2002, pp.89). This InterPARES statement refers to records, which are of diminished value if they not authentic. We can widen this discussion to include other kinds of digital materials as well. The UNESCO Guidelines consider the attributes that contribute to authenticity as the elements that give material its value . . . the selection process should consider what those elements and characteristics are (UNESCO, 2003, p.75). These vary according to the kinds of material. For example, for e-mail messages it may be decided that users of a large collection of electronic mail messages only need to see elements that can be characterised as content information, such as the name and address of the sender, subject, date and time, recipients, and the message, in a standardised structure with only the most simple of formatting (UNESCO, 2003, p.78).
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
For other materials where the structures are much less well-dened, different elements and characteristics will be considered of value. Earlier in this chapter it was noted that we need to identify the equivalents of content, structure, and context for digital materials that are not records. Can this content, structure, and context notion assist us to develop more precise preservation selection criteria for application to digital materials? To test this we might use three examples: videotapes maintained by a government-owned broadcasting companys library, web sites, and dissertations and theses in digital form. Note that these examples are described here only as a kind of proof of concept to suggest that this approach will assist the development of viable selection criteria. The rst example is not strictly of digital material but some of the same principles apply. For videotapes maintained by a government-owned broadcasting companys library, the information content is signicant, as much material of national heritage signicance is produced by this broadcaster. Because the videotapes exist in a number of formats, technological obsolescence of recording and playback equipment is a major issue. Selection criteria should consider the obsolescence rating for each format, such as lower risk for formats such as VHS, vulnerable for Digital Betacam (Sony), threatened for 1-inch Umatic, and critically endangered for 1-inch SMPTE (Ampex). For this material the structure (record form) becomes a crucial criterion for selection. The context in which these videotapes were created also becomes important to consider: the broadcasting company has a legal deposit obligation for some material it produces, and for other material it has negotiated rights with intellectual property owners. For web sites (the second example) we need to document the structure (record form). This will be html, perhaps with Java applets, CGI, Cold Fusion and other related applications or add-ons. There is a need to preserve documentation about software. Some web sites may need to be retained for legal purposes. For the third example, dissertations and theses in digital form, we need to document the structure (record form), which may be word-processing and/or imaging software. Contextual information to retain could include documentation that explains the reasons why and the conditions under which this material was created. These examples are summarized in Figure 4.2.
Structure (record form)
Content (information)
Context (linkages) Legal deposit regulations; intellectual property rights of other parties Legal requirements Regulations under which submitted; value as a record of the universitys activities
Videotapes VHS, Digital Betacam Signicant national (Sony), 1-inch Umatic, heritage content 1-inch SMPTE (e.g. documentaries) (Ampex) Web sites Digital theses Html, Java applets, etc Word-processing software, pdf Signicant intellectual content Signicant intellectual content
Figure 4.2 Content, Structure and Context of Some Digital Materials
72
Selection of materials for preservation is a uid eld, in which many variables are constantly being redened and more thinking needs to be done. Digital materials add extra complications. Codication of these variables into frameworks that can be applied in practice is being attempted. One attempt is the appraisal toolkit for electronic records of the Public Records Ofce in the United Kingdom (Public Record Ofce, 2000).
How much to select?

The discussions about how much to select, and about the feasibility of selecting all digital materials for preservation (that is, applying little or no selection), have been largely carried out in relation to web archiving. In this context the opposing ends of the spectrum are the selective approach exemplied by the National Library of Australias PANDORA, and the whole domain harvesting approach of the Internet Archive. Both have their strong proponents, and for web sites the concept of targeted acquisitions supplemented by periodic trawls is gaining ground. The cases for and against selection in the digital preservation context are presented in this way: Advocates of a comprehensive approach argue that any information may turn out to have long-term value, and that the costs of detailed selection are greater than the costs of collecting and storing everything. Advocates of a more selective approach argue that it allows them to create collections of high value resources, with some assurance of technical quality and an opportunity to negotiate access rights with producers. There may well be a place for both approaches, as they are likely to produce quite different collections of digital heritage that are valued for different purposes (UNESCO, 2003, p.76). Other advocates of a selective approach suggest that sampling may prove useful. Vogt-OConnor, for example, includes sampling in her three approaches to selection. Statistical sampling methods avoid human judgment and bias, and have the merit of providing a non-editorial sketch of the whole, although material of enduring value will be missed. Traditional selection criteria, with their particular values and belief systems, can be applied. Or a combination can be applied: statistically sample everything, to an extent that is economically feasible . . . [then] select any items that t the collecting statement of the archives and which are evaluated as having high value . . . [and nally] pray for technological help and funding from the government, individuals, and the organizations beneting most from the new technologies, the software and hardware rms (Vogt-OConnor, 1999, p.23). We must also be aware of the fact that notions of signicance will vary over time. In a discussion of material of national signicance, Lyall suggests, as do other writers about selection in the digital environment, that national signicance cannot always be allocated to material at the time of its acquisition. The importance of an item or collection may change over time (Lyall, 1993, p.74). Variation in signicance over time becomes difcult to manage in the digital environment, because, as indicated above, it is necessary to select at the very start of the life of digital materials. Perhaps this argues for a greater emphasis
Conclusion 73
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
on sampling as a method to select a representative body of digital objects or collections. Whatever point on the selection spectrum is chosen, and at whatever point on the timeline, a cautious approach is urged by the UNESCO Guidelines because a decision not to preserve is usually a nal one for digital materials (UNESCO, 2003, p.75). A triage system might be used, so that material denitely worth preserving and material denitely not worth preserving are identied, and interim preservation measures are applied to the rest until a denite decision about selection can be made.
Conclusion
While it is reasonably straightforward to identify the criteria traditionally used by libraries when they select for preservation, and to describe appraisal practice for archival material, they cannot be applied directly to the selection of digital materials for preservation. A combination of existing criteria, weighted differently, with new criteria is needed. As one example, Eastwood suggests that the cost of preservation of digital materials will become rather more determinative of the outcome of appraisal than it was for traditional materials (Eastwood, 2003). Further research into selection is needed, as it is into other parts of the life cycle of digital materials. Such investigation needs to take into account the high levels of human input currently required in selection decisions and their cost. By identifying selection processes that can be automated, it may be possible to reduce the level of resources needed for this aspect of digital preservation. The outcomes may well result in radically different approaches (NSF-DELOS Working Group on Digital Archiving and Preservation, 2003, p.7). It would be easy to become obsessed by the challenges of selection, but consider this: preservation . . . is a relative rather than an absolute concept because objects change over time as do our approaches to viewing or interpreting those objects (Cloonan, 2001, p.235). So we should be reconciled to never getting it completely right for all time and to the frequently changing concepts in the future. Nevertheless, it is our responsibility to rise to the challenges posed with some understanding of the consequences of ignoring them.
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Chapter 5
What Attributes of Digital Materials Do We Preserve?
Introduction
There is an inherent paradox in digital preservation. On the one hand, it aims to deliver the past to the future in an unaltered, authentic state. On the other hand, doing so inevitably requires some alteration (Thibodeau, 2002, p.28) What parts of digital objects do we need to keep into the future? Is it the essence of a document? Should we be able to render the digital original in its native environment? (Van der Werf, 2002) In 1915 the fossil remains of an ancient Pleistocene hominid were discovered in the Piltdown quarries in Sussex, following the discovery of other fossil remains from this quarry in the preceding three years. The discovery of this Piltdown man transformed thinking about the evolution of mankind. However, in the 1950s it was clearly established that Piltdown man was a case of outright deliberate fraud, made up of fragments from a wide geographical and temporal range of sources (Harter, 1999). There is a real possibility that we might see the same issues arising with digital information as arose with Piltdown man, if digital materials are reconstructed, interpreted and used for evidential purposes without appropriate knowledge of their context, leading, at the very least, to erroneous conclusions, and perhaps even to fraud. A repeat of the Piltdown man saga is denitely a possibility, perhaps even a probability it may even have occurred already in the digital context. The characteristics of digital materials predispose them to such misinterpretation and fraud. Preserving digital materials inevitably means altering them, because the processes and techniques we apply require changes to be made. This raises Thibodeaus inherent paradox, quoted at the head of this chapter. We need, therefore, to determine the level of change that is acceptable, so that we can claim to have preserved digital materials in a state that is close to the original.
76
We also need to have a clear idea about what level of loss of digital data is acceptable to us. We need to understand better the nature of the changes that we must make in order to preserve digital materials. As well as gaining a deeper understanding of the changes that we make, we need to know more about exactly what it is we are trying to preserve. What characteristics, attributes, essential elements, signicant properties many terms are used of digital materials do we seek to retain access to? What is the essence of digital materials? Whereas this question was a simple one to answer in the non-digital context, where typically we sought to conserve and preserve the original artifact, for digital materials it is not so straightforward. The example of e-mails, referred to in the previous chapter, is commonly used to illustrate this point. The essential elements of these are relatively easy to identify: the content information (senders name and address, subject, time, date, recipient(s), text of the message itself) and a simple standardized structure (UNESCO, 2003, p.78). But for other digital materials the essential elements are considerably less standardized, if they are standardized at all, so identifying what attributes of these we need to retain is difcult. The UNESCO Guidelines note, perhaps rather simplistically: some materials are much more difcult to characterise, and expectations about how they will be re-presented for use, especially to an open-ended community of potential users, may be so hard to dene in advance that it becomes almost impossible (UNESCO, 2003, p.78). A good example to illustrate this difculty comes from the realm of the web. Some web sites do not have a clearly-dened audience, so it becomes more complex to determine what users of these sites might wish to see in the future. It is, therefore, harder to dene those characteristics of web sites that it is essential to preserve. Many web sites are dynamic and interactive, with search and retrieval aspects intrinsically bound with content. Therefore we need to preserve the functionality of the search and retrieval components, as well as the data that the functionality interacts with to generate the required output (Smith, 2003, p.1). In order to determine the essential elements of digital materials we want to retain access to, we need to know in the future that they are what they purport to be, that they are complete and have not been altered or corrupted (Ross, 2002, p.7). If we cannot ensure this, we will be victim to the frauds exemplied by Piltdown man. Considerable energy has been expended, particularly in the recordkeeping context, on determining what authenticity and integrity mean in the digital environment. If these characteristics cannot be established, then the genuineness of digital materials and our ability to use them for evidential purposes are devalued. How do we demonstrate that digital materials have not been altered? What actions do we have to take to establish authenticity and integrity and maintain them over time, so that the future user can be condent that the digital materials they are using retain their original meaning? (Ross, 2002, p.7). Another concept that we need to know more about when we address the question what attributes of digital materials do we preserve? is that of acceptable loss. It is unlikely, if not impossible, that we can preserve all of the attributes and functionality of digital materials, but we do not know much at all about
Digital materials, technology and data 77
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
what levels of loss are acceptable to users. In the words of a statement about research needs in digital preservation, how can we measure what loss is acceptable? What tools can be developed to inform future users about the relationship between the original digital entity and what they receive in response to a query? (NSF-DELOS Working Group on Digital Archiving and Preservation, 2003, pp.1920). Current knowledge of the attributes of digital materials and of how important they are for users is insufcient to preserve digital materials effectively. We need to know more: If an object is preserved only to enhance access and use, some transformations might be desirable that would be prohibited if they transformed attributes deemed essential to the object . . . It would seem obvious that strategies, tactics and method for the preservation of digital objects ought to be informed by a rich understanding of their nature and the specic objectives for preserving them in digital form. The nature of digital objects denes what we need to preserve (Thibodeau, 1999). This chapter examines some of these issues.
Digital materials, technology and data

Digital materials are the result of the mediation of information technology and data. That is, the digital object that is developed on a particular software and hardware platform can be viewed as it was originally conceived only on that particular platform. Viewing web sites illustrates this point. The software used to view them the web browsers are congurable by the user in relation to characteristics such as the font size and colour, unless the web site designer insists that the browser displays a site in a particular way. The experience of viewing that web site will be different for users who access that site using software and hardware that differs from that used to develop the site. Strictly speaking, observes Eastwood, it is not possible to preserve an electronic record. It is only possible to preserve the ability to reproduce an electronic record. It is always necessary to receive from storage the binary digits that make up the record and process them through some software for delivery or presentation (Eastwood, 2002, p.77, citing Thibodeau, 2000). This mediation aspect the rendering or presentation (and re-presentation) of the digital les is an important point to appreciate for digital preservation. Digital objects are not simply a bit-stream; rather, they are a combination of hardware, software and computer le. The bit-streams of computer les have to be processed technically before they make sense to a user:
Hardware
Software
Computer le
Digital object
78
The computer le may look and behave differently from how it was originally intended to look and behave, if the hardware and software combination differs. This difference may be so extreme that the authenticity of the digital object is called into question (Digital Preservation Testbed, 2003, pp.1314). (These concepts have been considerably extended in the consideration of digital objects as physical, logical or conceptual objects in Chapter 1.) Some implications of this dependency of digital objects on software are listed in the UNESCO Guidelines. Three different kinds of software dependency are identied: digital objects that are simple and less dependent on specic software (plain text in ASCII les is an example) digital objects that depend on more complex software that is, however, generic (such as HTML) and digital objects that depend on software specic to a particular hardware platform or operating environment (such as spreadsheets and word processing software). The example of web sites that incorporate search and retrieval functions indicates that some digital objects contain executable les and software programs, and others may contain combinations of all of the above. The extent to which the digital object is dependent on software has a signicant effect on the way in which it can be preserved (UNESCO, 2003, p.124). Furthermore, it is not possible to preserve originals of digital materials, because migration is probably inevitable at some stage in the preservation process, and migration involves technology. We are preserving copies of the originals, preservation copies made according to the particular methods and strategies that are appropriate or expedient (Gilliland-Swetland, 2002, p.198). Crucial to the development of digital preservation was the realization that separating the technology from the data was the key. The technology (hardware and software) was changing rapidly and would continue to do so. Earlier thinking placed a high emphasis on preserving the ability to read computer les in the software that was used to create them. Digital preservation solutions such as museums of technology and emulation were proposed. Once it was acknowledged that the digital le would have to be transferred from one medium to another and that it could be viewed using a hardware/software combination different from that with which it was created, the focus shifted from maintaining access to obsolete software and hardware to determining which characteristics or attributes of the digital le are essential to maintain access to in the future, and how this can be achieved on current and future hardware/software combinations. However, this change in thinking brought its own problems, as an early statement about electronic records recognized: The content of a traditional record is recorded on a medium (storage device, like a piece of paper) and cannot be separated from this medium. The content of an electronic record is also recorded on a medium, but from time to time it has to be separated from the original device and transferred to other, and often different types of, storage devices whenever it is retrieved or when necessitated by technological obsolescence. Unlike traditional records, an electronic record is therefore not
Digital materials, technology and data 79
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
permanently attached to a specic medium or storage device, so the opportunities for corruption grow. This presents additional problems in ensuring that the records authenticity and reliability are maintained (International Council on Archives. Committee on Electronic Records, 1997, p.22). The history of the response to digital preservation of the archives community in Australia illustrates this changing awareness and focus. Early responses to the issues of preserving electronic records, dating from around 1990, stemmed from pre-digital preservation attitudes, and tended towards the requirement that the artifacts themselves be preserved and kept accessible. Consequently emphasis was placed on maintaining hardware and on the longevity of media types such as disks and tapes. By the mid-1990s the thinking had changed and it was accepted, according to National Archives of Australia staff in May and June 2004, that the solution had something to do with software applications . . . the archives community generally hovered around this idea that somehow or other, if you could preserve the application, then you could preserve the data. However, by the late 1990s the archives community recognized that the problem is a data problem, its not a software application problem. A new understanding of how software interacts with data was emerging: At the end of the day records are just data and data are just records, and in that process we came more to a clearer understanding of how software applications interact with data to produce records . . . That radically changed the possibilities of nding solutions, and we found the solution very quickly . . . We really started to give serious attention to this in the year 2000 and we had our basic approach worked out within a year. The basic idea [was] of normalizing materials into XML structures and ensuring that those XML structures have appropriate metadata kept with them, so that . . . the provenance and all the other authenticity issues about them could be re-established (Interviews with National Archives of Australia staff, May-June 2004). Once we have realized that the key is to ensure the preservation of the data, regardless of the hardware and software platforms that the data will be processed on, then we need to pay a lot more attention to that data to the attributes of the data to be preserved and the level of acceptable loss and alteration of the data. As already noted, the process of preserving digital materials inevitably means that they are altered. While at one level digital preservation is a relatively simple matter of preserving the bit-stream, the result is meaningless unless the means to read and interpret that bit-stream are available, either by preserving them or developing software that makes them accessible. We might, for example, emulate a software application with a resulting loss of some functionality, affecting the ability of that software to fully interpret the data. What effect does this have on the authenticity of the digital object? Is some loss acceptable? If we can determine what attributes of that digital object are essential for us to understand it, then we can seek to preserve these attributes. For example, the colours presented on the screen may not be signicant for text content, but it could be signicant for understanding a web page. Document-like digital
80
materials, typically in PDF, HTML, XML, or SGML le formats, embody data content, structure and presentation (Van der Werf, 2002) and can be read using software other than that in which they were created. (In fact, this is the very characteristic of these le formats that makes them so useful for digital preservation purposes.) Other digital materials, typically educational software, games, and web animations, are executable les, that is, they are in a format that is directly read by a computer as a program and is run (executed). These are usually in proprietary le formats, not re-usable in other contexts (Van der Werf, 2002). The determination of which attributes of digital objects are essential to maintain into the future has been addressed largely by the recordkeeping community in relation to establishing the requirements for authentic digital records (considered later in this chapter). Once these have been established, combinations of data, software and hardware that will re-present those elements as accurately as required need to be maintained (UNESCO, 2003, pp.123124). An allied question is: What level of loss or alteration of data is acceptable? The UNESCO Guidelines suggest that decisions about this are likely to be required by digital preservation programs; as already noted, some loss of content, loss of the original look and feel, or loss of some original functions is likely. Further thinking about this question will assist in determining what the essential attributes of digital objects are (UNESCO, 2003, p.123).
The importance of preserving context

In addition to determining the essential attributes of digital materials and developing and applying techniques to maintain them over time, other information about digital materials must be identied and maintained over time, such as information about the context in which the data was created and existed. This concept is derived from recordkeeping theory and practice. Weber states: A good archive contains not only the end document, for instance the text of a document, but also documents which allow you to reconstruct the process which led to the agreement (Kasteren, 2002, p.9, quoting Weber). Gilliland-Swetland points out the value that is placed on context in standard archival practice: When materials are generated by the activity of an individual or organization, an interdependent relationship exists between the materials and their creator. A complex web of relationships also exists between the materials and the historical, legal, and procedural contexts of their development as well as among all materials created by the same activity. The organic nature of records refers to all these interrelationships, and archival practices are designed to collectively document, capture, and exploit them. These practices recognize that the value of an individual record is derived in part from the sequence of records within which it is located. They also recognize that it can be difcult to understand an individual record without understanding its historical, legal, procedural, and documentary aspect (Gilliland-Swetland, 2000, pp.16,18). She uses the example of a web page without its hyperlinks to illustrate this point: it may be less valuable to users because of its diminished evidential context.
The OAIS reference model 81
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
The requirement to trace the process of creation and change is doubly significant for digital materials, which, as we have seen, are altered in the process of preserving them. It is signicant because the authenticity of digital objects whether we believe that they are what they claim to be, our ability to trust them as accurate depends on knowing how and when, and by whom, they have been altered. In other words, we need to preserve information about the context of the data (this is also noted in Chapter 4). For records (that is, records in the recordkeeping context, records as evidence of transactions) to be authentic, knowledge about the business context and about the interrelationship between records is required. We should be able to identify why, when, where, by whom and so on they were created (or received) and used as indicators of the records trustworthiness. Put another way, it should be possible to position a record in the time and context in which it purports to originate (Hofman, 2002, pp.1819).
The OAIS reference model

The Open Archival Information System (OAIS) reference model brings together many of the concepts discussed so far in this chapter and points to others that are relevant to the question of which attributes of digital materials we are attempting to maintain into the future. This model was developed to provide a common framework for describing and comparing architectures and operations of digital archives. It is being adopted throughout the world as an important design schema for digital preservation systems. The digital preservation community is adopting the concepts and terminology embedded within it. The OAIS model is emerging as the rst international standard in digital preservation (Beagrie, 2003, p.5). The OAIS reference model was developed by the Consultative Committee on Space Data (based at NASA in the USA) and is now an international standard ISO 14721:2003. It is based on the concept of open standards, increasingly being recognized as a stable foundation for digital preservation activities. As Breeding explains, one of the most important digital preservation strategies involves taking advantage of open standards whenever appropriate and avoiding proprietary data formats when possible (Breeding, 2002, p.48). The OAIS model has been used as the basis of digital preservation systems, such as those developed by OCLC, the Koninklijke Bibliotheeks e-Depot, and the CEDARS archiving demonstrator project, more details of which are provided in Chapter 9. Part of the signicance of the OAIS reference model lies in the establishment of a common language for discussion of digital preservation: its Foreword notes that it establishes a common framework of terms and concepts to allow existing and future archives to be more meaningfully compared and contrasted and to promote standardization (Consultative Committee for Space Data Systems, 2002, p.iii). For example, it denes long term preservation: long enough to be concerned with the impact of changing technologies, including support for new media and data formats, or with a changing user community. Long Term may extend indefinitely (Consultative Committee for Space Data Systems, 2002, p.11). It makes a clear distinction between simple data storage and long-term preservation: A major purpose of this reference model is to facilitate a much wider understanding of what is required to preserve and access information
82
for the Long Term. To avoid confusion with simple bit storage functions, the reference model denes an Open Archival Information System (OAIS) which performs a long-term information preservation and access function (Consultative Committee for Space Data Systems, 2002, p.21). The signicance of the OAIS reference model also lies in its articulation of the functional requirements of a digital archival system. It denes ve functions: the Ingest function responsible for receiving information from producers and preparing it for storage and management within the archive the Archival storage function which handles the storage, maintenance and retrieval of the AIPs [Archive Information Packages] held by the archive the Data Management function which coordinates the Descriptive Information pertaining to the archives AIPs, in addition to system information used in support of the archives function the Administration function which manages the day-to-day operation of the archive the Access function which helps consumers to identify and obtain descriptions of relevant information in the archive, and delivers information from the archive to consumers. These ve functions taken together . . . identify the key processes endemic to most systems dedicated to preserving digital information (Lavoie, 2000, pp.2829). They have proved to be inuential, one indication being the widespread adoption of the term ingest in digital preservation discussion. Of greatest relevance to this chapter, however, is the OAIS concept of information package, explained as a conceptual container of two types of information called Content Information and Preservation Description Information (PDI). The Content Information and PDI are viewed as being encapsulated and identiable by the Packaging Information. The resulting package is viewed as being discoverable by virtue of the Descriptive Information (Consultative Committee for Space Data Systems, 2002, p.25). In other words, a digital object consists of much more than just the content that we wish to preserve (the information which is the original target of preservation). It also comprises information that tells us what we need in order to preserve it (to ensure it is clearly identied, and to understand the environment in which the Content Information was created), information about its attributes (such as le formats), and so on. What is this information? And why is it an essential component of the digital materials package in relation to preservation? This information metadata is an essential part of the OAIS reference model and is increasingly considered vital to all digital preservation activities. OAIS is not straightforward to comprehend. Beagrie suggests that it needs to be further developed to make it more accessible, for example by producing accessible guides to major concepts behind the standard (Beagrie, 2003, p.7).
The role of metadata 83
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
The role of metadata

Metadata (structured information that describes, explains, locates, or otherwise makes it easier to retrieve, use, or manage an information resource (NISO, 2004, p.1)) has captured the attention of many. A considerable amount of energy has been channeled into its development, and many words have been written about it. A subset of this activity is concerned with preservation metadata. Preservation metadata is now considered an integral part of the strategies required for longterm maintenance of and access to digital materials, although some (such as Chivers and Feather, 1998, and Lazinger, 2001) claim for it an even more dominant role, suggesting it to be almost synonymous with digital preservation. Day notes that metadata is an intrinsic part of the current key digital preservation strategies of emulation, migration, and encapsulation (Day, 2004, p.255). The UNESCO Guidelines reinforce the essential nature of preservation metadata, listing among their fundamental principles four relating to metadata: 17. Digital heritage materials must be uniquely identied, and described using appropriate metadata for resource discovery, management and preservation. 18. Taking the right action later depends on adequate documentation. It is easier to document the characteristics of digital resources close to their source than it is to build that documentation later. 19. Preservation programmes should use standardised metadata schemas as they become available, for interoperability between programmes. 20. The links between digital objects and their metadata must be securely maintained, and the metadata must be preserved (UNESCO, 2004, p.23). The second of these principles is reinforced by the view of US experts that best practice for preservation metadata is to create it at the information creation stage (Meeting of Experts on Digital Preservation, 2004). The importance of metadata is corroborated by Australian digital preservation specialists interviewed in 2004. One was a self-confessed metaphile. Another considered that intellectual control was the key to effective digital preservation: the only way to manage [digital preservation] is to have an intellectual control system that applies across all media, that is not media-specic . . . building systems that are not technologically or system dependent, [where] the intellectual control system is a sort of conceptual system above whatever technology you use. From his experience of 10 or 15 years of work in the digital preservation environment, he claims that metadata works really really well, you know . . . [If] you use that across all technologies or media or record types, then you dont ever have a problem nding anything. Preservation metadata dened on the PADI web site as structured ways to describe and record information needed to manage the preservation of digital resources should not be confused with descriptive metadata, those based on schemas such as Dublin Core or AGLS (the Australian Government Locator Service). These have the primary aim of resource discovery, that is of enabling information resources to be identied and linked to users requests, and are best considered as a set of signposts for digital surfers . . . there to guide people to
84
resources (Wilson, 2003, p.33, citing Tom Baker). By comparison, preservation metadata stores technical details on the format, structure and use of the digital content, the history of all actions performed on the resource including changes and decisions, the authenticity information such as technical features or custody history, and the responsibilities and rights information applicable to preservation actions (Preservation metadata, 2003). Preservation metadata consists of two types of metadata: content information, and preservation description information (to use the terms of the OAIS reference model). Content information consists of details about the technical nature of the object which tells the system how to re-present the data as specic data types and formats. Preservation description information is other information needed for long-term management and use of the object, including identiers and bibliographic details, information on ownership and rights, provenance, history, context including relationships to other objects, and validation information (UNESCO, 2003, p.98). More specically, preservation metadata Identies the material for which a preservation programme has responsibility Communicates what is needed to maintain and protect data Communicates what is needed to re-present the intended object (or its dened essential elements) to a user when needed, regardless of changes in storage and access technologies Records the history and the effects of what happens to the object Documents the identity and integrity of the object as a basis for authenticity Allows a user and the preservation programme to understand the context of the object in storage and in use (UNESCO, 2003, p.97). There are, as yet, no commonly accepted standards for preservation metadata, although several contenders are in development. Perhaps the best-known schemes to date include those of the OCLC/RLG Working Group, the National Library of New Zealand, and the National Library of Australia; another Australian scheme is the National Archives of Australias recordkeeping metadata scheme. Early preservation metadata schemes were developed by NEDLIB and CEDARS, and the National Library of Australia. The National Library of Australias metadata scheme was grounded in that Librarys experience with digital preservation, including the PANDORA web archiving project, and was issued in 1999 (Lupovici, 2001, pp.67; National Library of Australia, 1999b; Bradley and Woodyard, 2000). An exposure draft issued by the National Library of Australia in October 1999 outlined the information it believed was needed to manage the preservation of its digital collections. This scheme was informed by other schemes and models, such as the OAIS Reference Model, the NEDLIB project, and activities relating to projects at the Library of Congress, CEDARS, the National Archives of Australia, and the Research Libraries Group, none of which was considered to be fully satisfactory for the National Library of Australias requirements at that time. The National Library of Australia model is meant to be a data output model, not a data input model; that is, it denes the information we want out of a metadata system, not necessarily what should be entered, how it should be entered, by whom and at what time; nor does it concern itself with how the metadata should be associated with what it is describing (National Library of Australia, 1999b). It consists of 25 elements:
The role of metadata 85
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Persistent Identier Date of Creation Structural Type Technical Infrastructure of Complex Object 5. File Description 6. Known System Requirements 7. Installation Requirements 8. Storage Information 9. Access Inhibitors 10. Finding and Searching Aids, and Access Facilitators 11. Preservation Action Permission 12. Validation 13. Relationships
1. 2. 3. 4.
14. 15. 16. 17. 18. 19. 20. 21. 22. 23.
Quirks Archiving Decision (work) Decision Reason (work) Institution Responsible for Archiving Decision (work) Archiving Decision (manifestation) Decision Reason (manifestation) Institution Responsible for Archiving Decision (manifestation) Intention Type Institution with preservation responsibility Process
24. Record Creator 25. Other.
Figure 5.1 Elements of the National Library of Australia Metadata Scheme
NEDLIB (the Networked European Deposit Library) published a metadata scheme in 2000, which was intended specically to manage the preservation of digital materials received through national deposit. CEDARS immediately developed the NEDLIBS metadata scheme further to cover administrative, technical and legal information (Lupovici, 2001, pp.67). These and other schemes were analysed by a Working Group on Preservation Metadata convened by OCLC and the Research Libraries Group (RLG), with representatives from the RLGs international membership, including Australia. This Working Group released a recommended preservation metadata framework in 2002 (OCLC/RLG Working Group on Preservation Metadata, 2002). The framework is based on the OAIS reference model recommendations. Following its release OCLC and RLG convened the PREMIS Working Group to focus on how preservation metadata could be implemented in digital preservation systems. In 2004 PREMIS released a report on metadata practices and other digital preservation activities of respondents in 13 countries (OCLC/RLG PREMIS Working Group, 2004). Among other outcomes will be an implementable set of core preservation metadata elements and a data dictionary (Guenther, 2004; Caplan, 2004). The National Library of New Zealand has been active in developing a metadata scheme for digital preservation as well as tools for extracting metadata automatically from digital materials. Its scheme is based on the OCLC/RLG metadata framework. Knight describes progress and expectations of the National Library of New Zealands preservation metadata scheme (Knight, 2003, p.18). Its metadata standards framework is a work in progress which, when complete, will provide a comprehensive statement of the standards environment within the National Library. The scheme has 18 elements describing the logical object, 13 elements to record the history of actions performed on the object, nine for
86
technical information about le types, and ve to record the history of changes made to the preservation metadata (Searle and Thompson, 2003). A resource discovery metadata scheme was released in 2000, followed by a preservation metadata scheme in November 2002, which was revised in June 2003. A data model and its associated XML schema denitions, released in July 2003, describe how the metadata scheme can be implemented. The metadata extraction tool developed by the National Library of New Zealand has attracted the attention of the international digital preservation community. It is designed to extract automatically preservation metadata from a range of le formats and output the extracted data in XML format so that it can be loaded into a preservation metadata repository. It is being developed to extract data from a number of common le formats, such as MS Word (various versions), WordPerfect, Open Ofce, MS Works, MS Excel, MS Powerpoint, TIFF, JPEG, WAV, MP3, HTML, PDF, GIF, and BMP. This tool is designed for use by the wider digital preservation community and it is hoped that any future development will be informed by that community, notes the National Library of New Zealands web site, describing the strongly collaborative approach taken by the international preservation metadata community. This metadata extraction tool, as well as information about it, is available free from the National Library of New Zealands web site. Its innovative nature and signicance was acknowledged by its short-listing for the digital preservation award, introduced in 2004 as one of the Pilgrim Trust Conservation Awards. The National Archives of Australias recordkeeping metadata scheme is part of its e-permanence initiative. It is designed to be used in conjunction with AGLS metadata, to uniquely identify records, authenticate them, document and preserve their content, context and structure over time, administer conditions of access and their disposal, track usage history and management, and simplify the transfer or migration of electronic records between computer systems (Robertson and Cunningham, 2000, p.196). Some of its elements are specic to preservation: no. 2, Rights Management, and no. 17, Preservation History (National Archives of Australia, 2001c). Although implementing preservation metadata undoubtedly has the potential to improve digital preservation practice, its full potential will not be realized until there is widespread agreement on a standard set of elements. US experts concluded that 107 elements were needed in a preservation metadata scheme (Meeting of Experts on Digital Preservation, 2004), and the current lack of consensus is apparent even from the brief descriptions of preservation metadata schemes, above. There is potential for reducing the costs of developing preservation metadata schemes, as well as reducing the costs of creating metadata by using automated metadata creation software. Standardization also allows greater sharing of information, making it easier to move digital objects from one archive to another, and encourages the standardization of preservation processes (UNESCO, 2003, pp.9293). Here, perhaps, is an analogy with the history of the MARC format in libraries: from a single format in the mid-1960s, it proliferated during the 1970s and 1980s, and only after several decades has it once again become a single internationally-adopted standard. The advantages of a limited number of standard schemes appear not to have yet been fully recognized by the preservation metadata community. Preservation metadata schemes continue to evolve. There is still relatively little experience of the effective application of preservation metadata. The
Persistent identiers 87
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
NSF-DELOS Working Group on Digital Archiving and Preservation notes that there has been limited evaluation of the effectiveness or cost of metadata for managing digital entities over time and suggests that research is needed, for example to demonstrate its value for specic purposes and to determine how much metadata is needed. Tools are needed to aid in the creation, authoring and management of metadata, such as those which automate, or partly automate, the creation of metadata; these are being developed, an example being the National Library of New Zealands metadata extraction tool. Also required are tools that manage metadata schemes so that they are useful over time, such as those to track the provenance of metadata schema, for version control, and to allow users to navigate from current metadata schema and ontologies to those used when the digital entity was created. The value of metadata could be assessed relative to the costs of extracting, creating and managing it, to provide a better understanding of the minimum amount of metadata necessary for digital preservation (NSF-DELOS Working Group on Digital Archiving and Preservation, 2003, p.19).
Persistent identiers
Persistent identiers are a form of metadata relating to digital materials on the web. The normal way of identifying web materials is to use a URL (Uniform Resource Locator). However, a URL points only to the one location of the material on the web, and, if the location of that material changes (for example, the material is moved from one domain name to another), the URL is altered and the material becomes inaccessible if the superceded URL is used. The persistence of URLs has been investigated in many studies (such as Koehler, 2004; Smith, 2004) and their tendency to change is widely recognized as an impediment to digital continuity. Persistent identiers have been developed to address this issue. Reliable identication of digital materials is essential for reliable long-term access and for ensuring their authenticity. Beagries survey of international preservation activities highlighted the need for persistent identiers, not just for web material but also for the linking and citation of primary research and data sets (Beagrie, 2003, p.7). Persistent identiers are, of course, not new: formal identier schemes such as ISBN, ISSN and ISMN have developed and been applied for many years in the pre-digital context. A persistent identier is a name for a resource which will remain the same regardless of where the resource is located (National Library of Australia, 2002b). If the material (or resource) is moved, the persistent identier provides access to it in its new location, provided that the persistent identier is maintained with the correct current associated location when the resource was moved (Persistent identiers, 2002). Persistent identiers can be considered as a form of preservation metadata, in the sub-set of preservation description information already noted above (UNESCO, 2003, p.98). Various kinds of persistent identiers have been developed and implemented. The Uniform Resource Name (URN) is a standard, persistent and unique identier for digital resources on the Internet (National Library of Australia, 2002b). URNs rely on a resolver service to link the URN to the location (URL) of the web site on which the digital material is currently hosted. The Persistent Uniform Resource Locator (PURL) is similar to a URL, but links to the resolution service
88
rather than the actual location of the digital material on the web. An example (hypothetical) of a PURL is http://purl.org.net/weibel/, where http:// is the protocol, purl.org.net/ is the resolver address, and weibel/ is the name. (This example is one of many given in answers to frequently asked questions about PURL (purl. oclc.org/docs/purl_faq.html#toc1.4).) PURLs were developed by OCLC as an interim system, awaiting the full implementation of URNs (National Library of Australia, 2002b). The Digital Object Identier (DOI), an implementation of the URN concept, was initiated by the Association of American Publishers to assist the publishing community with electronic commerce and copyright management of digital objects published on the Internet and allow the allocation of a unique digital identier to commercial digital publications (National Library of Australia, 2002b). It has achieved NISO standard status, as ANSI/ NISO Z39.84. DOIs can be applied to any kind of digital materials. The DOI system is managed by the International DOI Foundation, and DOIs are allocated by DOI Registration Agencies in the USA, Australasia, and Europe (International DOI Foundation, 2004?). Other persistent identiers have been developed, such as the Handle System (similar to URNs) and ARKs (Archival Resource Keys). Dack (2001) and the PADI web site (Persistent identiers, 2002) provide further information about them.
Authenticity
As noted in Chapter 1, denitions of terms taken from the pre-digital preservation paradigm are not always appropriate when applied to digital preservation. Discussions about digital preservation are further confused by the difference in terminology among librarians, recordkeepers and other interested groups. Terms like provenance, archiving, context, records, etc. are used with slightly different meanings . . . any discussion about preservation is challenged by confusion of terminology (Hofman, 2002, p.15). The term authenticity is one of these, and therefore we rst need to establish what is meant by it Authenticity is dened variously: the quality of genuineness and trustworthiness of some digital materials, as being what they purport to be, either as an original object or as a reliable copy derived by fully documented processes from an original (UNESCO, 2003, p.161) the digital material is what it purports to be (Jones and Beagrie, 2001, p.9) the degree of condence a user can have that the object is the same as that expected based on a prior reference or that it is what it purports to be (Authenticity, 1999) a document that is the same as that which a user expected based on a prior reference (Kerry, 2001, p.7, citing the CEDARS Project). Terms related to authenticity are: integrity, which for digital material is the state of being whole, uncorrupted and free of unauthorised and undocumented changes (UNESCO, 2003, p.162) essential elements or signicant properties, the elements, characteristics and attributes of a given digital object that must be preserved in order to represent its essential meaning or purpose (UNESCO, 2003, pp.161162); and
Authenticity
89
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
identity, the attributes of a digital object that uniquely characterize and distinguish it from other digital objects (Duranti, 2003, p.2). The UNESCO Guidelines get to the heart of the matter, stating succinctly that Authenticity derives from being able to trust both the identity of an object that it is what it says it is, and has not been confused with some other object and the integrity of the object that it has not been changed in ways that change its meaning and that evaluating, maintaining and providing evidence of continued authenticity are key responsibilities for most preservation programmes (UNESCO, 2003, p.113). Authenticity is valued in a number of different contexts for a variety of reasons. It is critical where digital materials are used as evidence. For records, the authenticity of the record is paramount: all who use it need to be able to trust that it is what it purports to be. Because regulated industries, nancial institutions, hospitals and clinics, and public entities are legally obliged to keep records over decades for purposes of accountability, continuity of operations and organizational memory their authenticity needs to be safeguarded (NSFDELOS Working Group on Digital Archiving and Preservation, 2003, p.4). Scholars will need to feel condent that references they cite will stay the same over time, and for participants in the legal system, there needs to be an assurance that material can withstand legal evidential requirements (Jones and Beagrie, 2001, p.25). The list goes on: heritage materials are valued because they are authentic; for scientic data, trust in their ongoing authenticity is critical, for without it they are of virtually no value (UNESCO, 2003, p.112). It could even be said that the future of government is at stake: the likely success of e-government initiatives will depend on their ability to demonstrate to the citizen that digital records are maintained over time in systems that guarantee their authenticity and integrity (NSF-DELOS Working Group on Digital Archiving and Preservation, 2003, p.4). The high value placed on authenticity has a long history. The Athenian government considered that a key function of the library should be to serve as a repository of trustworthy copies, at least for the works of Aeschylus, Sophocles, and Euripides. Their plays were considered to be superior to those of their successors and revivals of these plays became the dramatic mainstay. However, according to Casson (2002) the actors in them took liberties with the text to the extent that Lycurgus (leader of Athens from 338 to 325 BC) directed that Written versions of the tragedies of [Aeschylus, Sophocles, and Euripides] are to be preserved in the records ofce, and the city clerk is to read them, for purposes of comparison, to the actors playing the roles, and they are not to depart from them. In other words, notes Casson, an authoritative version of each play was to be kept on le, and the actors were to follow it, under penalty of law (Casson, 2002, pp.2930). Authenticity of non-digital records has traditionally been assured by the need for creators to rely upon their own active records, the xity of these records, a
90
documented unbroken chain of custody from the creators to the archivists, and the description of the archival record within a nding aid (Gilliland-Swetland, 2005). Authenticity of non-digital materials was determined by considering provenance, integrity and context. Artifacts are of value because their originality, delity, xity, and stability are preserved (Task Force on the Artifact in Library Collections, 2001, pp.1011). For museum objects, provenance (the chain of ownership and context of use of an object ((Signicance), 2001, p.37) is especially important. An artwork that has a doubtful provenance is diminished in value (in this case nancial); material from an archaeological site whose provenance has not been recorded is similarly reduced in irs reliability as evidence of past societies, and, therefore, in the uses to which it can be put. Integrity an objects condition, intactness and integrity provides an object with authenticity in that it is complete and in original condition ((Signicance), 2001, p.43). These determinations of authenticity and integrity are very closely linked to context. We need to be sure that the context in which the object was produced, discovered, owned, and stored is recorded. But this is more difcult to ensure for digital materials, and their authenticity, therefore, is harder to establish and maintain. Why is establishing and maintaining the authenticity of digital materials different from traditional materials? Why is it of such concern to those involved in digital preservation? As noted at the start of this chapter, preserving digital materials inevitably means altering them, because the processes and techniques we apply require changes to be made, changes that are easy to make. In the words of Jones and Beagrie (2001, p.9), condence in the authenticity of digital materials over time is particularly crucial owing to the ease with which alterations can be made. The threats to authenticity are intrinsic to the preservation processes for digital materials. These threats can be summarized as of three kinds: 1. the possibility of a multiplicity of versions of a particular document (Authenticity 1999) resulting from confusion in identifying data, changes to identiers, or failure to document the relationships between different versions or copies (UNESCO, 2003, p.113) 2. changes caused by preservation processes, for example, the common practice of migrating information from one system or carrier to another may result in changes, as may adding metadata, creating new copies and other processes (UNESCO, 2003, p.113) 3. changes to the content of the material (UNESCO, 2003, p.113). In addition there are, as noted in Chapters 2 and 3, threats that apply to all digital data what the UNESCO Guidelines call the ongoing integrity of data and they affect authenticity. Among the threats are breakdown of carriers, malicious acts, such as attacks by hackers or viruses, terrorist attacks, war, civil unrest (especially as they compromise power supplies and the integrity of buildings), accidental acts by staff, and res, oods and other natural disasters, and business failure (UNESCO, 2003, p.113). Because authenticity is so highly valued, digital preservation programs need to take appropriate steps to ensure that it is not compromised for the materials in their custody. The strategies applied to ensure the authenticity of digital materials include unique identiers, encapsulation techniques (the packaging together of the digital object, its metadata, and other associated data), digital
Authenticity
91
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
watermarking, digital signatures, encryption, digital time stamping, audit trails, controlled custody, and trusted repositories (Authenticity, 1999; NSF-DELOS Working Group on Digital Archiving and Preservation, 2003, p.6).
Signicant properties
Ensuring authenticity does not require keeping the materials in their original form without change: this is, as noted, impossible for digital materials, and is indeed inimical to digital preservation processes. We need to identify the attributes of digital materials that we must preserve to ensure authenticity, and also the attributes we do not need to preserve, because preservation processes often require transforming the original bitstream (NSF-DELOS Working Group on Digital Archiving and Preservation, 2003, p.17). This has been variously described as essence, core activities, signicant properties, essential elements, and performance. The term preservable essence of digital materials comes from the CEDARS Project (Kerry, 2001, p.7) and has been widely adopted. At the National Archives of Australia it is used to refer to the essential characteristics of a record (Heslop, Davis and Wilson, 2002, p.13). It is noted further below in this chapter and also in Chapter 8. Hofman discusses the term signicant properties in relation to records in digital form: One of the basic notions in a digital environment is the difference between what is shown on the screen and what is stored on the disk or medium. There isnt even always nor does there have to be a 1:1relationship between them . . . It is the message or the intellectual content which the author or creator intended to convey that has to be preserved. That content has a context, form and structure and in some cases also behaviour (e.g. spreadsheets) . . . in general we want to preserve the intellectual expression (object, if you like), not the digital components of which it consists per se, because, if one thing is obvious in a digital world, it is that fact that these digital components will change over time . . . What we want to achieve . . . is that in the future we will still be able to see, read and understand the documents or other information entities that were once produced for a certain purpose and in a certain context. In trying to achieve this, we of course need to preserve these digital components, but, as information technology will evolve, these components have to be migrated or in some cases emulated to be usable on future hard- and software platforms. Therefore, one of the issues is to identify what the intellectual aspects are. The emerging notion of Signicant Properties seems to acknowledge this (Hofman, 2002, p.17). Hedstrom and Lee dene signicant properties as those properties of digital objects that affect their quality, usability, rendering, and behaviour (2002, p.218). Microlming of paper documents preserve the signicant properties of most of these documents, that is their information content. But what is the equivalent for digital materials, which are considerably more complex? Hedstrom and Lee (2002) attempted to provide an empirical model for assisting decision-making about which signicant properties of digital materials need to be preserved.
92
The National Library of Australias Digital Preservation Issues Group has also investigated signicant properties, dening them in relation to four categories of desired functionality of digital objects that it wishes to preserve: 1) full dynamic/interactive functionality, 2) look/sound and feel, 3) intellectual content, and 4) description of what was (e.g. metadata for an object that no longer exists). Both the National Librarys experience and that of Hedstrom and Lee suggest that further research is required to ascertain signicant properties for specic types of digital materials in relation to their use by particular communities (Hedstrom and Lee, 2002, p.223). The idea of access to digital materials being analogous to a performance is gaining currency. Thibodeau tells us: It is necessary to recognize that, strictly speaking, it is not possible to preserve an electronic record. It is only possible to preserve the ability to reproduce an electronic record. It is always necessary to retrieve from storage the binary digits that make up the record and process them through some software for delivery or presentation. Analogously, a musical score does not actually store music. It stores a symbolic notation which, when processed by a musician on a suitable instrument, can produce music. Presuming the process is the right process and it is executed correctly, it is the output of such processing that is the record, not the stored bits that are subject to processing (Thibodeau, 2000, p.1). It is described in the UNESCO Guidelines. The idea is simple; the application of the same software and hardware to the same digital materials should create a presentation or performance that is the same every time. For preserved digital materials, their essential elements are presented during a performance sometime in the future: copying data from carrier to carrier, and providing the right tools to recreate the intended performance will preserve continuity of access to most digital objects. This apparently simple model is, however, rather more complex in practice: it may be hard to dene the performance that must be re-presented; it is usually difcult to work out what tools are needed once the original ones have been lost; the tools themselves typically rely on other tools that also may have been superseded; and it may be difcult to nd tools that will create the required performance in a reliable, cost-effective and timely way, especially in the context of many thousands, millions or more of digital objects. Despite such underlying complexities, the performance model helps in recognising what digital preservation programmes must aim for: the best means of re-presenting what users needs to access (UNESCO, 2003, p.36). An Australian digital preservation specialist interviewed in 2004 noted the National Archives of Australias view of performance and essence: issues like look and feel . . . are signicantly less important, because we can demonstrate to you that if we show you the same Word document on a different operating system, it can actually look different. If we show it on a different machine which has different fonts installed, it will look
Authenticity
93
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
different. If I alter the page setup, or even on different machines with the same page setup, youll get a different pagination, and if youve got an automatic footer with a page number in it, it will automatically change that pagination for you, and it wont be apparent to you that the pagination has changed. So a lot of that we have actually dened as being ephemeral or circumstantial aspects of the performance of the record in a particular situation . . . is not particularly relevant. I suppose if we got down to a very ne detail we might at certain points sort of say, well we have to say that because this record is a very important record and it matches lets say a hard copy record, then it really needs to be viewed with this sort of pagination and that sort of thing, but generally we dont get down to that level of detail and we would regard that as an extremely exceptional sort of thing to have to do. Weve dened most of these essence characteristics, in practice we dene them by dening the schema for the particular elements. What this boils down to is that for the materials being preserved we need to understand the characteristics that they embody, and the minimum set of these characteristics that need to be maintained in order to recreate the materials in the future. As well as needing to preserve the physical object (the physical form that carries the bit-stream) it is necessary to preserve the logical object (the computer readable code) and the conceptual object (the performance presented to a user (UNESCO, 2003, p.36)) that is meaningful to the human user. However, we also need to envisage the digital object as bundles of essential elements that embody the message, purpose, or features for which the material was chosen for preservation (UNESCO, 2003, p.36). Not all of the elements that make up a digital object are equally important in recreating the conceptual object. It is also necessary to understand the needs of the community of users for whom digital materials are being preserved. Community can be as closely dened as a specic organization or discipline group, or as extensive as the general public. Their needs determine the kind of material selected for preservation and the levels of authenticity required. Essence, signicant properties, or essential elements of the materials selected are dened in relation to the communitys requirements. For example, if the community is one where value is placed on the authenticity of records of transactions, that is their evidential value is high, then a lot of attention must be paid to maintaining the integrity of those records: ensuring that any alterations made to them are carried out only by authorized personnel and are appropriately documented, or that the records are preserved in an unalterable (read-only) form (UNESCO, 2003, p.77). However, some communities will not require that authenticity be proved to this extent. Ultimately, preservation programmes must decide how much to invest in ensuring that the authenticity of material in their care can be trusted, bearing in mind that object identity and data integrity are fundamental responsibilities (UNESCO, 2003, p.114). An Australian example of a communitys response to the challenges of authenticity can be seen in the Authenticated Electronic Editions project. This addresses the needs of textual editing scholars who require maintenance of the integrity of the core text while it is being proliferated, translated across platforms, manipulated, supplemented and analysed. To achieve this, the project has developed JITM (Just In Time Markup) (Australian Scholarly Editions Centre, 2001).
94
The UNESCO Guidelines suggest some questions to be posed in deciding on essential elements to preserve:
For whom should this material be kept? Do they have specic expectations about what they will be able to do with the material when it is re-presented? Why are the materials worth keeping? What gives them the value that warrants the trouble of preserving them? Is that value associated with evidence, information, artistic or aesthetic factors, signicant innovation, historic or cultural association, what a user can make the material do or do with the material, culturally signicant characteristics? Is the value tied to the way the material looks? (Would it be lost or signicantly degraded if the material looked different?) Is the value tied to the way the object works? (Would it be lost if particular functions were removed? Or if particular functions happened at a different speed or required different keystrokes?) Is the value tied to the context of the material? (Would it be lost if links embedded in the material did not work? Or if a user could no longer see evidence that connected the material with its original context?) Is it possible to distinguish between elements within each of these areas? For example, would advertising banners be considered an essential part of the way the material looked? Would some navigation elements or display functions be needed but not others? If it is difcult to dene what needs to be maintained, it may be easier to consider the impact of an element not being maintained, and to look for functions or elements that are denitely not needed.
Figure 5.2 Deciding on Essential Elements (From UNESCO, 2003, pp.7879)
Further elaboration of the example of e-mails, used in the previous chapter and earlier in this chapter, serves to illustrate some points about preserving authenticity. Authenticity in this case ensures the trustworthiness of the e-mail as a record of a transaction. An authentic record is one that can be proven a) to be what it purports to be, b) to have been created or sent to the person purported to have created or sent it, and c) to have been created or sent at the time purported (Millar, 2004, p.4). For an e-mail to be considered an authentic record, it must be demonstrated that it has not been changed or corrupted, and that it was created by a particular person. This is achieved by describing and preserving the original context of the records and by maintaining a chain of unbroken custody (Digital Preservation Testbed, 2003, p.15). Changes are acceptable as long as they do not alter the original meaning of the e-mail. E-mails can be considered as consisting of ve characteristics: context (the environment in which the digital record is made) content (the body of the record, regardless of structure, colour, position or font) structure (the structure as it was originally made and reproduced on the screen)
Research into authenticity 95
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
appearance (the nal presentation, what the records looks like when it appears on the screen) behaviour (the interactive characteristics of a records; that which enables us to manipulate and use the record so that new and extra content is displayed) (Digital Preservation Testbed, 2003, pp.1416). For e-mails, however, it could well be decided that it is only the content information that users require the name and address of the sender, subject, date and time, recipients, and the message, in a standardised structure with only the most simple of formatting (UNESCO, 2003, p.77). Clausen provides another example, of digital objects harvested from the web. He identies ve aspects of preservation quality readability, comprehensibility, appearance, functionality, and look and feel and gives examples of how these might be applied to a range of digital objects commonly encountered on the web (Clausen, 2004, pp.810). A third example, provided by an Australian digital preservation specialist interviewed in 2004, was of the Perseus web site (www.perseus.tufts.edu). He commented that the way the site is laid out, the way things relate to each other, I would have thought was sufciently signicant to warrant preservation apart from the actual content. The difculty of one aspect of this decision was noted: quite how one generalizes from that into a wider set of principles is a harder thing to try to come to terms with. More research into authenticity and its requirements in the digital world is required, although establishing these will be no easy task, as the authors of a CLIR publication emphasize (Authenticity in a Digital Environment, 2000). In the opinion of the NSF-DELOS Working Group on Digital Archiving and Preservation, the digital preservation research agenda includes research into tools that allow future users of digital materials to determine whether they are authentic (NSF-DELOS Working Group on Digital Archiving and Preservation, 2003, p.vii). We need a working denition of authenticity, closer denition of signicant properties and a clearer understanding of how these signicant properties affect use and access of digital materials, knowledge of how much change is acceptable before the authenticity of digital materials is compromised, among other things (Kerry, 2001, p.7). We also need strong legislative, organizational, and policy frameworks for managing digital records, and stronger technical and operational standards (Millar, 2004, p.8).
Research into authenticity

Several research projects have been inuential in determining requirements for digital materials to be considered as authentic. It is signicant that these projects originate in the recordkeeping professions and, consequently, are principally concerned with electronic records. It is not clear how some of their conclusions might relate to other kinds of digital materials. There have not as yet been any similar projects that have focused on other kinds of digital materials.
Functional Requirements for Evidence in Recordkeeping Project (Pittsburgh)

One of the ironies of digital preservation is that the web site on which the reports of the Functional Requirements for Evidence in Recordkeeping Project were to
96
be found was an early victim. The project was administered by the University of Pittsburgh. The web site hosting the working les of this project was deleted; however, the site can be accessed through the Internet Archive (Cox, 2000?). The project (commonly referred to as the Pittsburgh Project) investigated during the early to mid 1990s the functional requirements necessary for electronic recordkeeping systems to meet the needs of archivists better. However, its focus soon changed from system requirements to the requirements of electronic records to be considered as evidence. It developed a denition of what constituted an electronic record which placed high importance on metadata (Heazlewood, 2000, p.176).
InterPARES 1 and 2
The InterPARES (International Research on Permanent Authentic Records in Electronic Systems) project was based on an earlier project headed by Luciana Duranti at the University of British Columbia in the mid 1990s, which sought to establish the systems requirements for electronic recordkeeping. Its rst phase, now referred to as InterPARES 1, ran from 1999 to 2001. Gilliland-Swetland (2002) provides a detailed description of InterPARES 1 and its outcomes. InterPARES 1 focused on the preservation of records that are no longer current and was mainly concerned with records generated in databases and document management systems. It asked What is required to prove the authenticity of electronic records? and analysed case studies of electronic records systems to develop statements of requirements to create and maintain stable electronic records and for reproducing copies of preserved electronic records (USInterPARES Project, 2002, p.5). Recommendations about authenticity include: Records should be created and maintained in a trusted recordkeeping system, and preserved via a trusted custodian, who is able to maintain them for the long term without alteration. The custodian of records should assess the authenticity of the records before transfer to the preserver . . . The assessment of authenticity needs to . . . establish the records identity (the distinguishing characteristics of the record) and its integrity (the records wholeness and soundness) [and] establish evidence that the records have not been inappropriately altered . . . Records creators . . . should [meet] . . . the following Benchmark Requirements in the creation, handling, and maintenance of active records: Maintain expression of record attributes (especially those relating to identity and integrity), and evidence of the following: Effectively implemented, creator-dened access privileges. Protective procedures to prevent loss or corruption of records procedures and to prevent media and technology deterioration. Documentary forms of records (according to requirements of juridical systems or of creators). Specic rules for authentication of records (i.e., which records are authenticated, by whom, and how). Procedures that identify an authoritative record (when multiple copies exist). Procedures for removal and transfer of relevant documentation (involving removal of records from the electronic system) . . .
Trusted digital repositories 97
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
The trusted custodian should be able to attest to the authenticity of copies of inactive or preserved electronic records by meeting the following Baseline Requirements: Maintain controls over records transfer, maintenance, and reproduction. Retain documentation of reproduction processes. Capture, as part of archival description, any changes the records have undergone since they were rst created (US-InterPARES Project, 2002, pp.78). Other recommendations relating to preservation stress the importance of metadata, for instance, the need to maintain information about the original form of the record and the methods needed to translate between the stored digital components and the copy of the record presented for use. Thorough documentation of the entire process of preservation is also recommended (US-InterPARES Project, 2002, pp.910). InterPARES 2 was planned to run from 2002 to 2006. It is addressing issues of authenticity, reliability and accuracy of records, but differs from InterPARES 1 in that it examines these characteristics of records throughout their life-cycle, not just after they have been determined to be worthy of long-term retention, and in that it focuses on dynamic records, rather than those generated in databases and document management systems. It is also different because it examines records in the artistic and scientic domains.
Trusted digital repositories

InterPARES 1 dened the role of a trusted custodian in ensuring that authentic records are available. The concept of trusted digital repositories, the roles and responsibilities of which were articulated by the inuential 1996 Task Force on Archiving of Digital Information, is similar. A number of trusted organizations which have the capability of storing, migrating and providing access to digital materials are crucial for effective digital preservation, and a process of certication is essential to ensure trust (Task Force on Archiving of Digital Information, 1996, p.37). However, it is only in recent years that the manner in which they might operate has become clearer. Trusted digital repositories would be national and, increasingly, international systems of digital repositories that are . . . responsible for the long-term access to the worlds . . . heritage in digital form (RLG/OCLC Working Group on Digital Archive Attributes, 2002, p.[i]). To reach this level of operation much needs to be done. The RLG/OCLC Working Group on Digital Archive Attributes listed the tasks as: 1. Develop a process for the certication of digital repositories. 2. Research and create tools to identify the signicant attributes of digital materials that must be preserved. 3. Research and develop models for cooperative repository networks and services. 4. Develop systems for the unique, persistent identication of digital objects that expressly support long-term preservation. 5. Investigate and disseminate information about the complex relationship between digital preservation and intellectual property rights.
98
6. Determine the technical strategies that best provide for continuing access. 7. Dene the minimal-level metadata required for long-term management and develop tools to automatically generate and/or extract as much of it as possible (RLG/OCLC Working Group on Digital Archive Attributes, 2002, p.[i]). Because content creators and the owners and users of information need to be able to trust digital repositories, certication is a vital aspect of their establishment and operation. How certication might be achieved has been the subject of recent discussion. Not only is certication required, but auditing may also be required to monitor ongoing trustworthiness.
Conclusion
Chapter 1 notes that to ensure that digital materials remain usable in the future, access to them is required and not simply access, but access to all qualities of authenticity, accuracy and functionality (Jones and Beagrie, 2001, p.9). Authenticity of digital material needs to be dened in relation to users requirements, and the elements of the materials that meet these requirements must also be clearly dened. It is also essential that we realize that some loss of elements and functionality is inevitable. Consequently the levels of acceptable loss must also be clearly described. An Australian digital preservation specialist interviewed in 2004 claimed that theres no absolute answer to the question what kind of compromise should we accept? He noted that in lm archiving the whole eld is fraught with compromise: when we see a Technicolor lm of the 40s and 50s, we are actually watching an Eastmancolor copy usually, and the colour rendition has already changed, because Eastmancolor just has different spectral characteristics . . . So if you are going to give people the actual experience, how purist do you get? . . . there just are compromises in every direction. Once we have dened what levels of loss are acceptable, we can focus on how to preserve the essential elements. The next chapter examines strategies and techniques for digital preservation.
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Chapter 6
Overview of Digital Preservation Strategies
Introduction
There is as yet no viable long-term strategy to ensure that digital information will be readable into the future (Rothenberg, 1999b, p.v) Maintaining access to digital resources over the long-term involves interdependent strategies for preservation in the short to medium term based on safeguarding storage media, content and documentation, and computer software and hardware; and strategies for long-term preservation to address the issues of software and hardware obsolescence (Jones and Beagrie, 2001, p.95) High-level models for persistent repositories indicate that digital archiving and long-term preservation is best handled by separating archival storage of bits (storage management) from data management, logical representations, and higher level services that can be built on top of a persistent storage architecture (Workshop on Research Challenges in Digital Archiving and Long-term Preservation, 2003, p.xiv) This chapter provides an overview of the range of principles, strategies and practices that we have available to us at present for digital preservation. It also examines the requirements for approaches that will remain viable and effective in the future. It attempts to put the approaches identied to date, both those in current use and those that are emerging, into a useful context so that they can be reected on. Such reection may point to clearer ways forward in the future. A distinction is made between principles, strategies and practices. Principles are general ways of thinking, usually at the conceptual level; for example, it is a principle to use standards where they exist. Strategies are more concrete plans to achieve a particular long-term aim, as plans designed to achieve a particular long-term aim; an example of a strategy might be to limit the number
100 Overview of Digital Preservation Strategies
of standard le formats that are to be maintained. Practices are even more specic. A practice is what you need to do to implement and maintain a strategy, for example, the application of the necessary technology to maintain the standard le formats selected. Practices encompass technologies the machinery and equipment based on the application of scientic knowledge for practical purposes (as dened in the Concise Oxford English Dictionary). The terminology is used in different ways through the literature. For instance, ERPANET (the Electronic Resource Preservation and Access Network) uses the word technology inclusively to cover all hardware and software as well as methods and procedures: technology are all means that serve the purpose of preserving digital objects for as long as it [sic] is needed (ERPANET, 2003, p.2). Others have developed typologies of the principles, strategies and practices. (Typology is used here in the sense of systematic grouping to help understanding of things being studied by identifying attributes or qualities among them that link them together.) These typologies range from limited lists with a small number of categories, such as that of Lim, Ramaiah and Pitt (2003), to more sophisticated versions, such as Rothenbergs typology (2003). These are examined in more detail later in this chapter. Why might dwelling on these typologies help us to address digital preservation concerns? The UNESCO Guidelines suggest that strategies are still evolving and that there is, as yet, no universally applicable and practical solution to the problem of technological obsolescence for digital materials (UNESCO, 2003, pp.122,124125). Alan Howell notes 16 strategies to ensure the long-term sustainability and accessibility of digital objects (Howell, 2004, pp.3132). There are many ways of characterizing the range of strategies. For example, we can distinguish between passive approaches (such as improving digital storage media, improving storage and handling practices) or active approaches (such as refreshing, migration, emulation, encapsulation, normalization, and replication). Research and implementation work proceeds into all of these and more, for example, replication (the LOCKSS (Lots of Copies Keep Stuff Safe) Project), digital mass storage systems, and trusted repositories. The contention of this chapter is that, because such a wide range of approaches, strategies and practices is in use or development, determining their desirable characteristics could guide us to more rapid development of viable mechanisms to address the significant threat of loss of digital information. Developing such typologies could indeed heighten our understanding. The UNESCO Guidelines identify some strategies that are likely to be viable in the long term: the use of standards for data encoding, structuring and description, emulation, and migration of data (UNESCO, 2003, pp.124125). The Guidelines also note some principles that lie behind current approaches to preserving digital materials (UNESCO, 2003, section 17.1112). Despite guidance such as this, the list of possible strategies remains long and bewildering (see Figure 6.1). From this a smaller number of strategies is likely to emerge. This chapter rst notes some of the history of the development of principles and strategies for digital preservation. It then notes strategies that are being applied. Finally, some existing typologies of approaches, strategies and practices are described, and a further typology is proposed.
Introduction
101
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Principle, Strategy, Practice Analogue backups Backup and restore Bit-stream copying Canonicalization Digital archaeology Durable/persistent media Emulation Encapsulation Improving handling and storage Mass storage Media transfer Metadata Migration
Also referred to in the literature as . . . Output to permanent paper or microlm, page image techniques; Save page-images of artifacts
Translate artifacts into standard or canonical forms Data recovery Improved storage media Technology emulation Creating a persistent object; Digital tablets Digital mass storage systems
Format migration; Normalization then migration; Software migration; Backwards compatibility and version migration
Persistent archives Persistent object preservation Policy development Refreshing Replication Standard data formats Standards Technology preservation Viewers Virtual machines Other (not readily categorized) Developing policy, specications and procedures Refreshing data; Medium refreshing Replication; Redundancy; Keeping multiple copies Normalization; XML; Restricting the range of formats to be managed; Long term formats Reliance on standards; Use of standards Preserve technology; Maintain old technology; Technology watch Viewers for obsolete formats Virtual machines; UVC (Universal Virtual Computer) approach Save source-code of rendering software (for future reverse-engineering); Extract and save core contents of artifacts; Software repositories; Web domain harvesting; Formalization (replace artifacts by formal descriptions of themselves); Data extraction and structuring
Figure 6.1 Digital Preservation Principles, Strategies and Practices

(From Kransch, 1998; Soete, 1997, p.7; Hendley, 1998, p.4; Bullock, 1999; Woodyard, 2000; Heazlewood, 2000, pp.180184; Howell, 2000, pp.133134; Van der Werf, 2002; Howell, 2001, pp.142146; Strang, 2003; Walton, 2003, p.5; Lim, Ramaiah and Pitt, 2003, p.121; Digital Preservation Testbed, 2003, p.29; Rothenberg, 2003; UNESCO, 2003, pp.126149; Smith, 2003, p.43; Beagrie, 2003, pp.1516; Kenney et al., 2003)
Historical overview
Chapter 1 explored the need for a new preservation paradigm in the digital world. It noted that the principles that lie behind preservation practice, and the preservation techniques themselves, need to change. Where pre-digital preservation paradigm practices are based on the preservation of artifacts, the new preservation paradigm requires different ways of thinking which are still not completely clear. Some are accepted and understood: for example, that there is a need to actively maintain digital information from the moment of its creation, and that there is likely to be greater emphasis on collaboration and support from a wider range of stakeholders. It is perhaps inevitable that the path to these new ways of thinking should have its beginnings in elements of pre-digital preservation paradigm, such as the strong emphasis on techniques and processes based on artifacts and copying. This resulted in early digital preservation efforts focusing on the storage media (developing more durable media is an example) and on copying onto more durable storage media (the printing of digital objects to paper is perhaps an extreme example of this). It is salutary to review the literature of digital preservation over the last decade and to see how thinking has changed. An awareness of this historical background helps us understand better the requirements for effective digital preservation strategies and practices. One can summarize from the sources used to construct Figure 6.1. Before 2000 the list of possible strategies and practices was short: migration, emulation, and technology preservation were the constants, with mention of refreshing, output to permanent paper or microlm, and digital tablets. From about 2000 others are added. Migration, emulation, and technology preservation remain as the primary strategies and technologies. The list expands, however, to include encapsulation, digital archaeology, standardized le formats, printing to paper and writing to lm, canonicalization, developing policy, keeping multiple copies, and developing preservation metadata. Currently the list, having expanded still further, also includes bit-stream copying, analogue backups, persistent object preservation, digital archaeology, reliance on standards, viewers, replacing artifacts by formal descriptions, data extraction and structuring, the Universal Virtual Computer approach, Web domain harvesting, mass storage systems, redundancy, software repositories, and technology watch. The list also includes the do nothing strategy. There is a growing recognition in the literature that combinations of strategies and practices, rather than a single approach, will be required. Pre-digital paradigm concepts may also have led to the focus of some current digital preservation activities on technical problems, without sufcient effort being directed to wider issues such as developing appropriate public policy. Research in digital preservation can be characterized along similar lines. Early research focused on two areas: preserving the bitstream and extending the useful life of the entity on which information was recorded (NSF-DELOS Working Group on Digital Archiving and Preservation, 2003, pp.1011). Research into media has attempted to develop more durable media. Research directed at preserving the bit-stream has examined issues such as media migration. Much of this research has been concentrated on three strategies: normalization of digital content into a few common formats; migration of data
Who is doing what? 103
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
from obsolete to current computing platforms and applications; and emulation of obsolete platforms on current computing platforms (NSF-DELOS Working Group on Digital Archiving and Preservation, 2003, p.12). There still appears to prevail a mindset that wills us to develop the killer application, a single technical solution (such as Rothenbergs championing of emulation, described below), although this way of thinking is weakening. Two key realizations have allowed us to move towards a new preservation paradigm and to develop new approaches, strategies and practices. One is the concept of separating digital content from the technologies; the other is the need to embrace a wider range of stakeholders in digital preservation processes and to accommodate them in all-embracing policies and guidelines. These are examined in the rest of this chapter and in the two following chapters.
Who is doing what?

One way of thinking about the applicability and long-term viability of specic strategies and practices is to consider those that are actually being applied, and see if there are lessons to be learned from observing current practice. Digital preservation is a eld in which, so far, there has been considerable invention but relatively little application. In 1998 Hedstrom reported that methods being applied fall far short of what is required to preserve digital materials. She noted that all preservation methods require trade-offs between what is desirable from the standpoint of functionality, dependability, and cost and what is possible and affordable and this could be observed in the methods being applied: printing to paper or microlm, with the loss of retrieval and reuse potential, and migration strategies, which by normalizing to a standardized data format often result in loss of information about document structures and relationships (Hedstrom, 1998, p.195). In 1999 Hedstrom and Montgomery reported the results of a survey of 54 Research Library Group members. Only 13 institutions had established digital preservation methods in place, the most common being transfer accessions to new media (25), refresh (18), migrate (17), limit formats accessioned (16) and standards for archival masters (7) (Hedstrom and Montgomery, 1999, pp.1415). By 2003 Walton could observe that, although there was no consensus on the best method or methods, it is worth noting that the 4 major national archives which either have existing digital preservation programs . . . or are beginning one . . . have chosen to use a strategy which emphasizes preserving objects over preserving technologies: conversion to standard format (NA, NARA, NAA) or format migration (PRO) (Walton, 2003, p.7). Also in 2003, Bryans small survey of United States manuscript repositories, which sought to ascertain their practice in managing preservation of born-digital material, noted that of the nine respondents, four print electronic documents to paper and three migrate them to a server (Bryan, 2003). The European Commission-funded ERPANET Project provides case studies of the experiences of and approaches to digital preservation in a range of sectors
and institutions. These note some general strategies, for example, migration, and reducing the proliferation of formats. These case studies make observations about the pharmaceutical, publishing and telecommunications sectors. In the pharmaceutical industry, size and proliferation of formats are the main obstacles to the preservation of objects. Migration to new data formats was carried out when necessary, PDF (Portable Data Format) had become the standard format for preservation purposes, and there was a general belief that digital preservation solutions should come from outside the industry. In the publishing industries PDF proved a very popular format alongside TIFF (Tagged Image File Format), XML (Extensible Markup Language), and SGML (Standardized General Markup Language) for distribution and preservation. The telecommunication organizations relied on migration carried out when their business software was updated (Ross, Greenan and McKinney, 2003). Two more extensive surveys published in 2004 provide rmer data. A survey of digital preservation practice in 21 natural science and scientic publishing organizations operating on an international level concluded that Migration remains the preservation strategy of choice; it is still too soon for most archives to have undergone a signicant technological change. Other than the large data archives, which have existed for many years, archives have not yet faced large-scale technological changes. This means that migration remains the strategy for most of the materials of interest to libraries, archives, and publishers (Hodge and Frangakis, 2004, p.2 italics as in the original). The strategies used in this community were Transformation to a Preservation Format (for example, ASCII and XML), migration, and Migration On-Request [where] the original version of the material is retained and when necessary, conversion tools are applied to convert the original to the format required by the user this had been tested but not applied (Hodge and Frangakis, 2004, pp.3840). An international survey of current digital preservation practice in national libraries, state libraries, university and research libraries and consortia, archives, museums, and other organizations from 13 countries was carried out under the auspices of OCLC and the Research Libraries Group and reported in 2004 (OCLC/RLG PREMIS Working Group, 2004). Of its 48 respondents, 92 per cent were implementing (or planned to implement) normalization, migration, or emulation, most indicating that they applied more than one strategy. The most popular strategies are bit-level preservation (refreshing), 85 per cent, then restricting submissions and normalization (both ways to limit the variety of formats that have to be managed), migration, and migration-on-demand, in that order. Ten per cent were using emulation (OCLC/RLG PREMIS Working Group, 2004, p.6). Other strategies included technology preservation, the UVC (Universal Virtual Computer), and a digital ontology, dened as a description of the structure of the digital entity (OCLC/RLG PREMIS Working Group, 2004, p.36). Three conclusions from this survey are of interest. The rst is that only one respondent in production checked a single strategy (OCLC/RLG PREMIS Working Group, 2004, p.36), reinforcing the point made earlier that combinations of strategies and technologies are considered to be the most effective approach. The second is that because there is an assumption that there
The Australian experience 105
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
will be additional technical options in the future, decisions about digital preservation are being made on a short-term basis (OCLC/RLG PREMIS Working Group, 2004, p.37). Finally, the report noted that less than a quarter of respondents had actual production experience in implementing active preservation strategies (OCLC/RLG PREMIS Working Group, 2004, p.53). This is a signicant point. It emphasizes that there is still not a lot of operational experience in digital preservation. This makes the experience of those who have applied digital preservation strategies and technologies in more than a test situation all the more important.
The Australian experience

The views of Australian digital preservation specialists noted in this book are derived from their extensive operational experience. In interviews of Australian digital preservation specialists during 2004 the question was posed What makes a good digital preservation strategy? Their answers can assist us to determine the characteristics of viable preservation strategies and practices, and guide us to more rapid development of viable strategies to address the signicant threat of the loss of digital information. Five themes emerged: 1) 2) 3) 4) 5) Societal and organizational missions Knowing what you are preserving Standards Operational matters Technical issues.
The rst theme (societal and organizational missions) involves concepts such as assured funding, a sustainable supportive environment over time, knowing the context in which preservation occurs, and community will. A sustainable environment that supports digital preservation over time is essential, and this is related to funding: You need to have your long-term funding sorted out. To be sustainable, digital preservation activities had to be built into normal operating activity and a good understanding of the context was required: the example was provided of the need to understand the business drivers within a specic corporate context, where digital preservation was not seen as core business. To secure funding for digital preservation in this context, the preservation manager has to link it into the drivers in that particular sector. To fully grapple with the problem . . . its very short-term work that were doing, there needs to be a reason . . . a will as a community to hang onto things. The second theme (knowing what you are preserving) was expressed in terms of the need to think clearly about exactly what it was that we are trying to preserve. For Australian preservation managers from the recordkeeping context this was expressed in terms of essence, the essential parts, the things about a record that we think are signicant enough to be retained which had to be dened in advance. These concepts were closely linked to the third theme (standards), which noted strategies such as metadata, data formats that remain accessible, normalization, and standard data formats. Most respondents commented on standards as an essential aspect of any preservation strategy: one noted that for starters you have to have very clearly dened standards and
that may include the diversity of standards for different types of material, and another suggested that its doing it to the highest standard that is appropriate to what youre doing. The importance of standards was emphasized by a member of the archives community who described why XML had been adopted. Its multiple attractions included its stability as a standard (its actually got a very good track record. Its been in existence for . . . over 25 years now) and its widespread and increasing use. Metadata was noted as another area in which standards were essential. However, most attention was paid to standard data formats as a digital preservation strategy. A respondent from the archives community noted that the core of our strategy is to preserve records in data formats that are going to remain accessible over time. And what that means is that we actually convert . . . we call it normalization . . . the source records, which are usually in proprietary data formats into our archival data format, which is non-proprietary . . . it is our intention to always have nonproprietary data formats as our archival format. At the moment its XML. A fourth group of points made during the interviews have been grouped together under the term operational aspects. These points can be summarized as: capture the material rst, build digital preservation into normal operations, integrate digital preservation processes fully, keep data moving. There is a need to capture the data rst: you cant preserve it unless youve got it. So thats trying to get the thing into the cycle of preservation in the rst place. The importance was stressed of ensuring that digital preservation was established as part of normal operating activity, so that it was considered as a mainstream activity and became part of the normal resourcing issues. Another respondent described this in terms of an integrated response: I would certainly use the word infrastructure . . . [a digital preservation strategy] is not something you can just put down in one line. Its a whole series of integrated responses. Another theme was that data needed to be kept alive: the only strategy that will work is one where there are open, in a sense living, digital networks and systems that are extremely robust and have high levels of redundancy built into them i.e. there is duplication all over the place, there is distributed responsibility, there is distributed access. Therefore the stuff has to be able to migrate . . . just to keep moving . . . so migration has to be part of the system. You know, theres no question about that. The fth theme was technical issues. Although there was relatively little comment in these interviews about technical issues needing to be taken into account in effective digital preservation strategies, one that was noted was the importance of not relying on proprietary data formats or systems. A respondent noted this as having . . . control systems that are non-system-dependent . . . so that they exist independently . . . the objects themselves have got to be in a non-proprietary format. According to the Australian specialists interviewed, an effective digital preservation strategy has these characteristics: It should be part of a sustainable environment that supports digital preservation over time The context in which digital preservation operates, and therefore in which the strategy is placed, needs to be clearly understood A very far-reaching forward plan should exist for it It should be built into normal operating activities
Criteria for effective strategies and practices 107
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
It should be based on clear denitions of what it is that we are trying to preserve, the essence of a record It should be based on stable, widely used, and clearly dened standards, including a limited number of standard data formats where possible It requires sufcient management information metadata and preservation metadata It should recognize that digital preservation is an active process It should not be based on proprietary data formats or systems.
Criteria for effective strategies and practices

It is also worth noting what other commentators have decided to be the requirements for effective digital preservation strategies and practices. One way of doing this is to analyze responses to the threats to digital preservation. Respondents to a survey of RLG members in 1999 ranked the threats, in order of signicance, as technology obsolescence, insufcient resources, insufcient planning; and physical condition of materials (Hedstrom and Montgomery, 1999, p.19). This suggests that effective approaches to and strategies for digital preservation need to address these concerns, so our list of criteria could include the requirements that the approach or strategy should be independent of technology, that it must be adequately resourced, that it is planned in detail, and that attention is paid to improving the physical condition of media. Jeff Rothenberg has commented extensively on some of the requirements. He encourages a sound technical approach as the basis of any approach, recognizing that there are many inter-related issues that must also be addressed, such as technical, administrative, procedural, organizational, and policy issues (Rothenberg, 1999b, p.6). He elaborates that any technical solution must be also able to cope with issues of corruption of information, privacy, authentication, validation, and preserving intellectual property rights and must be feasible in terms of the societal and institutional responsibilities and the costs required to implement it (Rothenberg, 1999b, p.9). His ideal approach would be a single solution that is long-lived and can be applied uniformly, automatically, and in synchrony (for example, at every future refresh cycle) to all types of documents and all media, with minimal human intervention (Rothenberg, 1999b, p.16): The long-term digital preservation problem calls for a long-lived solution that does not require continual heroic effort or repeated invention of new approaches . . . This approach must be extensible, since we cannot predict future changes, and it must not require labor-intensive translation or examination of individual documents. It must handle current and future documents of unknown type in a uniform way, while being capable of evolving as necessary. Furthermore, it should allow exible choices and tradeoffs among priorities such as access, delity, and ease of document management (Rothenberg, 1999b, p.30). This is a tall order. Implicit in Rothenbergs comments is the idea that there is a single technical solution he is well-known as championing emulation as this single solution; but others do not agree. More recently, Lavoie and Dempsey have suggested that factors such as cost, user preferences, nature of the material,
whether it exists in multiple forms need to be taken into account when ascertaining which preservation strategies are appropriate (Lavoie and Dempsey, 2004). Thibodeau has identied four criteria by which to ascertain the best strategy: its feasibility (is there software and/or hardware capable of doing it?), its sustainability (can it be done into the future, or can an alternative future path can be identied?), its practicality (can it be applied within reasonable limits of difculty and expense?), and its appropriateness (this criterion relating to the type of objects and why we are preserving them) (Thibodeau, 2002, pp.1516). To illustrate these he describes a spectrum of possibilities, ranging from preserving technology itself to preserving objects that were produced using information technology (IT) (see Figure 6.3). The suitability of any strategy and practice, which Thibodeau calls methods, can be determined only with respect to the nature of the material being preserved. Computer games, for example, are much more likely to require methods that fall at the preserve technology end of this spectrum, whereas e-mails fall at the preserve object end. Some methods are general and some apply only to specic technologies, for example specic software and/or hardware, or specic data types. Therefore, the range of applicability is another basis for evaluating preservation methods (Thibodeau, 2002, p.17). Nor must we forget the requirements for authenticity, noted in Chapter 5, in digital preservation. Because an authentic digital object is one whose genuineness can be assumed on the basis of one or more of the following: mode, form, state of transmission, and manner of preservation and custody (Ross, 2002, p.7), it follows that effective digital preservation strategies and technologies must ensure that these requirements are not forgotten. We should keep in mind Gilliland-Swetlands cautionary comments: Counterintuitively, perhaps, it is during the preservation of digital materials that evidential value is often most at risk of being compromised. Digital preservation techniques have moved beyond a concern for the longevity of digital media to a concern for the preservation of the information stored in those media during recurrent migration to new software and hardware. In the process, many of the intrinsic characteristics of information objects can disappear data structures can be modied and presentation of the object on a computer screen can be altered (Gilliland-Swetland, 2000, pp.1112). ERPANET has produced a guide to assist with selecting digital preservation technologies in its erpaTool series (ERPANET, 2003). They provide a list of evaluation factors, reproduced here, with minor amendments, as Figure 6.2. Finally, we need to note the effect of national information infrastructures on approaches and strategies for digital preservation. Effective approaches and strategies will work best if supported within these infrastructures. Also relevant here are the possibilities for cooperation: for example, Beagries 2002 survey of national digital preservation activities indicated that national libraries offered
Factors
Remarks
General
Maturity Experience Spread
Standardization; open specications
It is important to include not only the price of system components, but all cost of implementing and maintaining the system. See for this purpose the erpaTool on cost orientation. Objects Legislation Are the objects subject to specic legislation which asks for a specic form, format, storage medium, or accessibility? Such regulations are basic conditions for the selection of technologies. Characteristics How can the main characteristics of the objects and their context be preserved without threatening authenticity and integrity? As conversion of digital objects for preservation is often difcult and costly, all selected technologies must be able to treat current and, as far as possible, also future objects. Consider also carefully what consequences a selected technology has regarding the creation or preparation of the objects for preservation. Preservation period As digital systems only have a life span of about ve to 10 years at highest and digital objects must be preserved for much longer periods, it is important that the system is able to efciently export objects and their context data in standard formats in order to migrate them into a new system. People Skills Does the selected technology need specic skills which must be available in-house? Are these skills already available or can they easily be acquired? Staff for maintenance Is the appropriately skilled work force in the right number for the maintenance available? Experience Is there sufcient experience with the technology for support in case of difculties available? (in-house or easily accessible in the region) Procedures Workow Can the technology easily be implemented in the preservation workow or can the workow be adapted to the technology without major difculties or loss of efciency? Flexibility Can the technology exibly be implemented? Does it allow changes in the preservation procedures? Good practices Are good practices in using the technology already established? Quality requirements Can the technology meet the previously dened quality standards? However, automatic or semi-automatic techniques are difcult to be applied for heterogeneous collections, complex objects, and high quality requirements.
Reliability Modularity and exibility Costs
Is the technology fully developed and are there already systems in productive use? Are there already veriable experiences in applying the technology for the preservation of similar objects? Is the technology widespread enough to guarantee that it will be supported by the manufacturers during the desired lifespan of the preservation system? Is the technology based on standards and are the specications of all the critical elements laid open by the manufacturers or at least deposited with an independent and trusted third party and available there in case of the dissolution or downfall of the manufacturers? Does the technology work reliably and can the reliability of the outcome easily be checked? Is it easily possible to add new components at low cost, to change or update them?
Figure 6.2 Factors to Consider when Selecting Digital Preservation Technologies (Modied from ERPANET, 2003, pp.45)
several recurring suggestions for international cooperation in developing effective strategies for long-term preservation (Beagrie, 2003, p.6), and the example of a preservation technology watch was offered.
Typologies of principles, strategies and practices

Figure 6.1 provides an unclassied list of most of the principles, strategies and practices for digital preservation noted in recent literature. It is worth attempting to categorize these, to develop a typology (or typologies) to assist us to understand them better, so that we can develop viable mechanisms more rapidly to address the signicant threat of the loss of digital information. Most of the literature lists, rather than categorizes, the strategies. Five publications, those of Jones and Beagrie (2001), Walton (2003), Thibodeau (2002), Rothenberg (2003) and the UNESCO Guidelines (2003), do more than simply list them and deserve further attention. Jones and Beagries inuential handbook Preservation Management of Digital Materials (2001) divides digital preservation strategies into primary and secondary strategies. Primary strategies are those which might be selected by an archiving repository for medium to long-term preservation of digital materials for which they have accepted responsibility. Secondary strategies are those which might be employed in the short to medium term either by the repository with long-term preservation responsibility and/or by those with a more transient interest in the materials (Jones and Beagrie, 2001, p.102). Primary preservation strategies are migration and emulation; secondary strategies are technology preservation, adherence to standards, backwards compatibility, encapsulation, permanent identiers, converting to stable analogue format, and digital archaeology. Jones and Beagrie note that sometimes secondary strategies will be employed as a holding action, to defer the need for primary strategies to be applied. They consider that there are two dominant strategies, migration and emulation, and that the need to apply these can on occasion be deferred and/or simplied if appropriate preventive preservation procedures such as storage and maintenance . . . and selected secondary preservation strategies are applied. They also note another strategy, converting to an analogue preservation format and its signicant limitations such as the loss of digital characteristics, and they characterize digital archaeology as a fallback position (Jones and Beagrie, 2001, p.103). The need to apply more than one strategy is indicated: employing a judicious mix of secondary strategies . . . with responsible storage and maintenance decisions . . . has the potential signicantly to reduce both risks of losing access to digital resources in the short-term and costs of preserving them in the long-term (Jones and Beagrie, 2001, p.103). Walton suggests ve categories: 1) Preserve old technology 2) Emulate old technology 3) Format migration
Typologies of principles, strategies and practices 111
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
4) Standardize formats 5) Persistent object preservation (Walton, 2003, pp.57). Walton discusses digital records within the context of the InterPARES Projects concern to establish the characteristics of authentic digital records. His ve categories fall into two types: those concerned primarily with the technology, and those primarily with the data formats. The two categories of strategies that focus primarily on the technology are preserve old technology and emulate old technology. Both of these leave the digital components of the record exactly as received, concentrating instead on the technology (software, operating system, hardware) that is required to reproduce the digital object. Walton suggests that the merit of this approach is that authenticity is much more likely to be maintained, but it is impractical because it requires multiple solutions (that is, one for every distinct category of equipment being preserved) rather than a single one, and writing emulator software is labour-intensive and costly. There is also, for the rst method, the major issue of technology obsolescence. Waltons three categories that focus primarily on the data formats are format migration, standardize formats, and persistent object preservation. Format migration attempts to ensure that the data remains live by continually updating it so it can be reproduced using current technology. Format standardization converts the data to a limited number of non-proprietary formats (for example, TIFF for images, XML for text) and there is thus no need to rely on one technology to read them, as many kinds of software can do this. Both of these categories minimize the need to rely on software and hardware that will become obsolete, but both require high levels of resources to manage ongoing migration projects or convert to standard formats. Waltons fth category, persistent object preservation, attempts to separate data completely from the technology. All aspects of the record are abstracted, or represented by other data, to the highest possible degree in order to render them independent of any specic technology. Digital materials are described in terms that are independent of software or hardware: an example is electronic records converted to XML and then encapsulated with metadata. Walton notes, however, that its operational viability remains to be demonstrated. (All quotes in the preceding two paragraphs come from Walton, 2003, pp.57.) Waltons approach is based on distinguishing between data preservation and maintaining the technology. Thibodeau extends this approach. He describes a preservation spectrum, which has at one end preserve technology, and at the other preserve objects. He suggests principles to guide us in positioning candidate methods across the spectrum: On the preserve technology end of the spectrum, methods that attempt to keep data in specic logical or physical formats and to use technology originally associated with those formats to access the data and reproduce the objects In the middle of the spectrum, methods that migrate data formats as technology changes, enabling use of state-of-the-art technology for discovery, access, and reproduction On the preserve objects end of the spectrum, methods that focus on preserving essential characteristics of objects that are dened explicitly and independently of specic hardware or software (Thibodeau, 2002).
But these are not sufcient: also to be taken into account is the suitability of a method for the material being preserved. Some methods are general, and others apply only to specic technologies. Thibodeau accordingly proposes that the methods suitability (which he designates as applicability) needs also to be taken account of. He has suggested a graph of preservation methods, reproduced here as Figure 6.3. Rothenberg takes a different tack. His classication is based on the extent to which strategies and practices are complete. His overview of proposed approaches to preservation (Rothenberg, 2003) has three categories: nonsolutions, partial solutions, and potentially complete solutions. Non-solutions include do nothing and digital archaeology. The longer list is of partial solutions: Save page-images of artifacts, Extract and save core contents of artifacts, Translate artifacts into standard or canonical forms (without migration), Rely on viewer programs to render obsolete formats in the future, Save metadata to help interpret saved bit streams (assisted archaeology), Save source-code of rendering software (for future reverse-engineering). It is Rothenbergs potentially complete solutions that we should pay most attention to. It has only three entries, suggesting that there is still signicant work to be done to develop a range of viable strategies and practices: Formalization (replace artifacts by formal descriptions of themselves) Migration (repeatedly convert artifacts into new artifacts) Emulation (run original rendering software on virtually recreated hardware) (Rothenberg, 2003). The categorization in the UNESCO Guidelines (2003) is different again, being based in part on the length of time that the strategies are likely to provide a
General Digital preservation methods
Universal virtual computer
Persistent archives Object interchange format
Applicability
Typed object conversion Virtual machine Emulation Rosetta Stone translation Format standardization
Re-engineer software
Programmable chips
Specific
Viewer Maintain original technology
Version migration
Preserve technology
Objective
Preserve objects
Figure 6.3 Digital Preservation Methods (From Thibodeau, 2002, p.19)
Typologies of principles, strategies and practices 113
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
viable result, and in part on what stage of the preservation process the principal effort is needed: Investment strategies (primarily involving investment of effort at the start): Use of standards Data extraction and structuring Encapsulation Restricting the range of formats to be managed UVC (Universal Virtual Computer) approach
Short-term strategies (likely to work best over the short-term only): Technology preservation Backwards compatibility and version migration Migration (which may also work over longer periods) Medium- to long-term strategies (likely to work over longer periods): (Migration) Viewers Emulation (UVC approach)
Alternative strategies: Non-digital approaches Data recovery Combinations (UNESCO, 2003, pp.126127). These Guidelines distinguish between general principles, and specic strategies and practices. For example, they suggest that the principles (which they label as strategies) for preserving digital materials include: Working with producers (creators and distributors) to apply standards Recognising that selection of material to be preserved is essential Placing the material in a safe place Using structured metadata and other documentation to control material, facilitate access and support preservation processes Protecting the integrity and identity of data Effectively managing preservation programs (UNESCO, 2003, p.37). Nine further specic principles on which current strategies and practices are based are articulated in the UNESCO Guidelines. These can be grouped into four categories: conversion to analogue form, maintaining the original format, standardizing formats, and migration. Conversion to analog form involves converting data to a human readable form on a stable carrier such as paper, lm or metal. Maintaining the original format is the most popular category, with four strategies: Making the data self-describing and self-sustaining by packaging it with metadata and with links to software that will continue to provide access for some time; Maintaining the data in its original form . . . and providing tools
that will re-present it . . . using the original software and hardware . . . or using new software that emulates the behaviour of the original software and/or hardware; Maintaining the data and providing new presentation software (viewers) that will render an acceptable presentation of it for each new operating environment; Providing specications for emulating the original means of access on a theoretical intermediate computer platform, as a bridge to later emulation in a future operating environment. Standardising formats includes either creating data in, or converting data to, a highly standardised form of encoding and/or document structure (or le format) that will continue to be widely recognised by computer systems for a long time, or converting the data to a format where the means of access will be easier to nd. Migration involves converting the data to new formats that can be read by each new technology; it also includes providing the ability for migration on demand by maintaining the data and recording enough information about it to allow a future user or manager to convert it to a then-readable form (UNESCO, 2003, p.125). Finally, preservation programmes, these guidelines suggest, must be organizationally viable and nancially sustainable, if they are to be effective and credible (UNESCO, 2003, pp.4344). This classication is noteworthy because of its emphasis on the combination of strategies and practices. In the absence of a single, universally applicable solution these guidelines support the application of multiple strategies in preservation programmes. These are likely to be based on the use of standards for data encoding, structuring and description that can be expected to remain recognisable for long periods; emulation of obsolete software or hardware in a new environment; and migration of data from one operating technology to another (UNESCO, 2003, pp.124126). The doubt expressed in the Guidelines as to the likelihood of a single solution ever being found is now generally shared. Whatever we might consider as mainstream approaches the consensus is that these are migration, technology preservation, emulation, and perhaps persistent object preservation they are not mutually exclusive, and will be deployed according to the kind of digital materials to which they are applied and their access requirements (Smith, 2003, p.43). Rothenberg is also aware of this, noting that For now, consider using multiple approaches in parallel: Digital archaeology (for records that are unlikely to be accessed), Page-image techniques (for simple records), Formal methods (when applicable and affordable), Standards (when available), Migration (when it needs to be done anyway, i.e. for active records; or as a stopgap), Emulation (if original behaviour is needed; or as a cheap backup, to preserve everything) (Rothenberg, 2003). None of the typologies described here are complete. They do not, for instance, note the principle of redundancy (multiple copies at multiple locations), best exemplied by the LOCKSS Project. However, they allow us to identify some of the principles that are important and to make some judgments about the applicability and viability in the longer term of specic strategies and practices. For example, they help us distinguish more clearly between preserve
A typology of digital preservation? 115
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
technology and preserve objects approaches, they encourage us to consider the nature of the digital object and the different requirements that each may have, with consequent different best strategies and technologies for their preservation, and they suggest that the tools we have available to us at present should be applied in combination.
A typology of digital preservation?

The desirable principles, strategies and practices that we need to have in place for digital preservation activities to be viable and effective can, perhaps, be listed, although the changing nature of the eld makes this task a challenging one. Such a list could include the aspects below, but it is essential to recognize that this list does not intend to be comprehensive: 1) Principles (general ways of thinking, usually at the conceptual level) a) Separate archival storage of bits from the processes of representing and managing those bits b) Actively maintain digital information continuously from the moment of its creation c) Give high priority to cooperative activities, including collaboration with and support from a wide range of stakeholders d) Combine strategies and technologies, rather than focus on a single solution e) Apply the principle of redundancy f) Ensure that digital preservation activities are part of a sustainable environment that supports digital preservation over time g) Build digital preservation into normal operating activities. 2) Desirable strategies (plans designed to achieve a particular long-term aim) a) Distinguish between medium- to long-term strategies (primary strategies) and short-term strategies (secondary strategies). (Short-term strategies focus on safeguarding storage media, content and documentation, and computer software and hardware; long-term strategies attempt to address software and hardware obsolescence.) b) Strategies should be feasible in terms of institutional responsibilities and the costs required to implement them c) Authenticity must be maintained, so strategies that emphasize preserving objects over preserving technologies are to be preferred d) Prefer strategies that recognize the nature of different kinds of digital materials and the different requirements that each may have. (Clear denitions of what it is that we are trying to preserve are therefore needed.) e) Adequate and appropriate management information metadata and preservation metadata are required f) Strategies should not be based on proprietary data formats or systems, but rather on stable, widely used and clearly dened standards g) Strategies should be designed around as limited a number of standard data formats as possible.
3) Desirable practices (the machinery and equipment based on the application of scientic knowledge for practical purposes) a) Practices should have the characteristics of feasibility (is there software and/or hardware capable of doing it?); sustainability (can it be done into the future, or can an alternative future path can be identied?); practicality (can it be applied within reasonable limits of difculty and expense?); appropriateness (this criterion relating to the type of objects and why we are preserving them) b) Distinguish between practices that leave the digital components of the record exactly as received (e.g. Preserve old technology, Emulate old technology), and those that focus on the data formats (e.g. format migration, standardize formats, and persistent object preservation).
Conclusion
Lavoie and Dempsey have noted that digital preservation is a set of agreed outcomes based on considerations such as the complexity of digital materials and the features that we collectively decide should be preserved. These considerations, they suggest, require that the choice of preservation strategy will need to reect a consensus of all stakeholders associated with the archived digital materials. Achieving such a consensus is difcult, and in some circumstances, impossible (Lavoie and Dempsey, 2004). How, then, are we to determine which strategies are best if we do not have this consensus to guide us? Lavoie and Dempsey continue: A second-best solution is for the digital repository to articulate clearly what outcomes can be expected from the preservation process. These outcomes should in turn be understood and validated by stakeholders. Communication between the repository and stakeholders, either to promote consensus on preservation outcomes, or for the repository to disclose and explain its preservation policies, mitigates the risk that the repositorys commitments are misaligned with stakeholder expectations (Lavoie and Dempsey, 2004). Clear communication with stakeholders is the key while we are in a rapidly evolving environment where there is little agreement about preferred strategies and practices, and where new approaches are being developed and trialed. The following two chapters are based on Thibodeaus typology (Figure 6.3), to which is added some aspects of Rothenbergs typology (noted above). They describe specic principles, strategies and practices: Chapter 7 examines approaches based on preserving the technology, and Chapter 8 examines approaches based on preserving the digital object.
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Chapter 7
Preserve Technology Approaches: Tried and Tested Methods
Introduction
The need to preserve digital materials runs counter to the market ethos of the computing industry, which requires high turn-over of hardware and software in order to survive nancially. This outlook necessitates rapid changes in formats and functionality and an unwillingness to support obsolete technology, all of which make it harder and harder to preserve access to digital materials (Heazlewood, 2002) There is, as yet, no universally applicable and practical solution to the problem of technological obsolescence for digital materials (UNESCO, 2003, p.124) The previous chapter presented a long and bewildering list of possible strategies that are available or under consideration for digital preservation (listed in Figure 6.1). It noted that they are still evolving and that no universally applicable and practical solution is available (UNESCO, 2003, pp.122,124). Chapter 6 also noted some of the characteristics of digital preservation strategies and practices and described some of the typologies proposed for their categorization. Chapters 7 and 8 take as their basis Thibodeaus typology, described in Chapter 6, add to it some parts of Rothenbergs typology, also described in Chapter 6, and use the result to describe specic principles, strategies and practices. The result is three categories: 1) Non-solutions (examined in this chapter) 2) Preserve Technology approaches (examined in this chapter) 3) Preserve Objects approaches (examined in Chapter 8). Non-solutions (the term used by Rothenberg (2003)) are:
118 Preserve Technology Approaches: Tried and Tested Methods
Do nothing Storage and handling practices Durable/persistent digital media Analogue backups Policy development Standards Digital archaeology.
The list of Preserve Technology approaches (the term used by Thibodeau (2002)) is much shorter: Technology preservation Emulation. The reader of this chapter and the next should keep rmly in mind that Preserve Technology and Preserve Objects, as noted in Chapter 6, are the two ends of a spectrum of possibilities, not discrete points on that spectrum. There are many points in between these two poles. The reader should also keep in mind that there are other ways of grouping the possible approaches. The grouping used in Chapters 7 and 8 represents the current state of evolution of digital preservation strategies and practices, which will change. For instance, many of the approaches described in this chapter technology preservation, adherence to standards, converting to stable analogue format, digital archaeology are what Jones and Beagrie refer to as secondary strategies, those which might be employed in the short to medium term either by the repository with long-term preservation responsibility and/or by those with a more transient interest in the materials (Jones and Beagrie, 2001, p.102). These strategies are, in a sense, holding actions, buying time for digital materials while longer-term strategies are developed, and practices that keep digital materials viable are implemented. But, as Jones and Beagrie make clear, these strategies and practices do not address the real issues of technology obsolescence; they only put off the need to make decisions. This chapter draws heavily upon two key sources: Preservation Management of Digital Materials: A Handbook by Maggie Jones and Neil Beagrie (2001), and the UNESCO Guidelines for the Preservation of Digital Heritage (2003). The reader may wish to update these by referring to PADI, the web portal maintained by the National Library of Australia with the support of the Council on Library and Information Resources, the Digital Preservation Coalition and ERPANET. PADI, which describes itself as a subject gateway to digital preservation resources, is an essential source for denitions and explanations of, and further references for, the strategies and practices covered in Chapters 7 and 8.
Non-solutions
Non-solutions is one of three categories suggested by Rothenberg (2003) in his overview of preservation approaches, the others being partial solutions, and potentially complete solutions. This categorization is useful because it emphasizes that some of the practices promoted as viable digital preservation techniques, both in the past and at present, are not likely to achieve the aims
Non-solutions
119
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
of most digital preservation activities that is, an assurance of long-term access to authentic digital materials. Rothenbergs examples of non-solutions are do nothing and digital archaeology. Leaving aside do nothing for the moment, these non-solutions are strategies and practices that are useful in the toolbox of approaches to digital preservation, or are essential parts of the infrastructure required for successful digital preservation, but they do not provide outcomes that result in preserved digital materials over time. Of the six non-solutions noted in this chapter, the rst (do nothing) is not a realistic option. Unlike non-digital materials, for which the principle of benign neglect can have validity, digital materials require active intervention right from the moment of their creation, if they are to survive. This is noted in more detail in Chapter 1. Two of the non-solutions (storage and handling practices, and durable/ persistent media) provide a breathing space, extending the life of digital storage media and, thereby, ensuring that the digital materials remain in good condition for longer. This potentially provides sufcient time to develop and implement strategies and practices that are viable over the long term. Storage and handling practices focus on environmental control, handling, building design, and security, although we note here only the rst two. The durable/persistent media non-solution is based on the premise that developing improved storage media media that will last longer, store more, and remain accessible for longer than current media will provide greater economies and efciencies by reducing the frequency of copying and the number of media items that are handled. However, both of these approaches are non-solutions because they focus on the media, avoiding the real issue of technological obsolescence. Proposing them as anything more than an interim solution is a classic example of the inappropriate application of pre-digital paradigm preservation thinking to digital materials. Another of the non-solutions, analogue backups, has the problem of negating the advantages of being digital, such as improved retrieval of the information content, ease of dissemination, and so on. (There is, of course, the possibility of converting back to digital form material that has been copied to analogue form, although this is obviously expensive and not likely to occur.) This approach is also referred to as analogue copying, output to permanent paper or microlm, and sometimes as page image techniques and saving page-images of artifacts, although these last two can also refer to making digital page images, such as PDF les. This approach, too, is an example of the inappropriate application of pre-digital paradigm preservation thinking to digital materials. Two more approaches included in the category of non-solutions (policy development, and standards) are in fact principles, following the denition and use of this term in Chapter 6, rather than strategies or practices. Attending to these provides a more favourable environment for successful digital preservation. Policy development is a prerequisite for digital preservation. Standards are also a prerequisite in that their development and application provides the same benets for digital preservation as they do for any cooperative endeavour. Standards are considered in another context, that of standard data formats, in Chapter 8. The nal non-solution, digital archaeology (also known as data recovery) is commonly recognized to be not precisely a preservation strategy (Jones and Beagrie, 2001, p.103) but, rather, a fall-back position. It encompasses actions taken to recover digital data that has become inaccessible using normal techniques, but where the value of the data warrants the time-consuming and expensive techniques of the data recovery specialist.
Do nothing
A discussion of the do nothing approach need not detain us long. As noted earlier in this chapter and in more detail in previous chapters, especially Chapter 1, it is not an option for digital materials. Doing nothing reduces to nil, in a very short period of time, the possibility of preserving digital materials. One familiar example is failure to monitor and respond to the deterioration of digital media (for example, a oppy disk) and the consequent inaccessibility of any data it might carry.
Storage and handling practices

Some of the reasons for media deterioration were noted in Chapter 3. This section notes techniques that are currently recommended as good practice for the storage and handling of media, in order to prolong the length of time that the data on them remains accessible. Application of these techniques assumes that the hardware and software required for these media is still available. (The sections Technology preservation and Emulation later in this chapter consider what is necessary to support this assumption.) Why is paying attention to the care and handing of digital media a good preservation strategy? The principal reason is that it reduces risk. Responsible storage and maintenance decisions have the potential signicantly to reduce both risks of losing access to digital resources in the short-term and costs of preserving them in the long-term (Jones and Beagrie, 2001, p.103). They include this approach in their list of secondary strategies, that is, strategies that are shortterm but, nonetheless, are worth committing resources to. Storing digital media in appropriate environmental conditions (this term refers to temperature, relative humidity, light and clean air levels) extends the time the media will last. It also protects them to some extent against accidental damage, as does appropriate handling. Of course, the need for attention to the storage and handling of digital media is not unique to them: it is fundamental to the preservation of other materials too. The National Library of Australias 1999 Draft Research Agenda for the Preservation of Physical Format Digital Publications lists what we already know, based on experience at the National Library of Australia and other institutions. One of its statements is particularly relevant to this section that magnetic media such as oppy disks have a relatively short useable life span, so we have to slow down the rate of deterioration and/or move what we want to preserve to a more stable carrier. The comment is made that we know quite a lot about the conditions of storage and handling that will maximise the useful life of these media (National Library of Australia, 1999a). Optimum environmental conditions for the storage of digital media are well documented. Figure 7.1 summarizes the recommendations for magnetic tape, CD-ROM, and DVDs. The life expectancy of digital storage media is determined by a number of factors. For optical disks, for example, it is a combination of the manufacturing quality, condition of the disk before recording, quality of the disk recording, handling and maintenance, [and] environmental conditions (Byers, 2003, p.12). Similar statements can be made about other media types. For all of these media one essential factor is to provide the optimum storage conditions, with temperature and relative humidity within the recommended ranges, and with as little
Non-solutions
121
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Media
Access storage (allows immediate access and playback) Temperature (C) Relative Humidity (%) 45 to 55 20 to 80 20 to 80 Room ambient (2575), maximum variation 20% 10 to 80
Long-term storage (preserves the media as long as possible) Temperature (C) 18 to 22 18 to 22 5 to 32 Relative Humidity (%) 35 to 45 35 to 45 20 to 60
Magnetic tape cassettes 12.7mm1 Magnetic tape cartridges1
18 to 24 10 to 45
Magnetic tape 4 & 5 to 45 8mm helical scan1 Magnetic tape3 Room ambient (15 to 23), maximum variation 4o 10 to 50
As low as 5 o, As low as 20%, maximum maximum variation 4o variation 10% 18 to 22 4 to 20 35 to 45 20 to 50
CD-ROM1 CD-ROM & DVD2

1 2
Jones and Beagrie, 2001, p.100, from BS 4783 Byers, 2003, p.vi 3 Based on Van Bogart, 1995, p.18.
Figure 7.1 Environmental Storage Conditions for Some Digital Media
uctuation as possible. Light levels are usually controlled; Byers notes that this is less of an issue for CD-ROMs and DVDs, but is important for CD-Rs that use dye-based technology, because light levels affect the rate at which the dye layer (the layer where the data is recorded) degrades (Byers, 2003, p.17). Another factor for media in long-term storage is the need to acclimatize the media if it is removed from this storage to room temperature for access (playback) purposes. Long-term storage areas are typically kept at much lower temperatures and relative humidity levels: for instance, Byers consensus of several reliable sources on the prudent care of CDs and DVDs (2003, p.1) recommends long-term storage levels as 18oC and 40% RH, but a lower temperature and RH . . . for extended-term storage (2003, p.vi). Careful handling of digital media also prolongs their life. For CDs and DVDs, for instance, careful handling minimizes the possibility of data loss resulting from scratching the data layer. Jones and Beagrie recommend that written guidelines and procedures for handling be produced (2001, p.101). For CDs and DVDs these procedures could include statements about handling the disks by their outer edge, returning them to storage cases straight after use, marking or labeling disks in a non-harmful way, not attempting to bend the disks, and cleaning them as seldom as possible, and then using approved methods only (Byers, 2003, pp.vi,19,2526). Figure 7.2 is a summary of recommendations about the storage and handling of digital media derived from several sources. It is not an exhaustive list.
Storage areas
Keep free of smoke, dust, dirt and other contaminants Store magnetic media away from strong magnetic elds Keep cool, dry, dust-free, stable and secure Minimize light levels, especially direct sunlight Prohibit smoking and eating in the storage area Minimize the threat from natural disasters Provide enclosures for media to afford additional protection Store in appropriate conditions any non-digital accompanying materials such as operating instructions and codebooks Monitor environmental conditions on a regular basis Monitor media condition on a systematic basis Handle with care Minimize the handling and use of archival media Establish guidelines and procedures for acclimatizing media if moving them from signicantly different storage conditions Do not place labels on optical disks and/or mark using a pen or pencil Follow recommended guidelines for labeling and marking Document the contents of the media, when created, and their frequency of use Choose digital media with preservation in mind: all digital media are not created equal (Howell, 2001, p.144) Use high quality equipment Keep access devices well maintained and clean Prepare for accidents
Monitoring Handling practices
Quality of media and equipment Disasters
Figure 7.2 Storage and Handling of Digital Media

(From Howell, 2001, p.144; Jones and Beagrie, 2001, pp.100,133; National Library of Australia, 2000?; Ross and Gow, 1999, p.43)
Durable/persistent digital media

Encouraging the manufacture and use of durable/persistent media is based on the idea that by using storage media that last longer and store more, the number of times copying of data is required and the number of media items to be handled is reduced, thereby producing economies through efciencies. This approach is a non-solution because it does not address technological obsolescence. It does, however, have a place among the practices used in digital preservation in that it buys time while other strategies and practices that will make long-term preservation viable are developed and implemented. Zwaneveld provides the example of CD-R disks and notes that an estimated 320 million units of CD-R disks will be produced in 1999, but that half of the manufacturers of this media are producing disks that will fail within ve years (Zwaneveld, 2000, p.2). One of the reasons for this is the economics of the media production industry; another
Non-solutions
123
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
is that the media are constantly evolving and production processes are in a state of almost constant evolution (Jones and Beagrie, 2001, p.129). Despite the limited use of these strategies and practices aimed at increasing the life and efciency of digital media, they are useful tools. The lifespan of digital storage media is well enough known for us to be aware of the risks we take when we commit digital materials to these media for preservation purposes. This awareness should make us wary of some of the claims made by the media manufacturers, which tend to reect the exuberance of scientists compounded by the hype of their marketing teams (Ross and Gow, 1999, p.iii). Attention to the selection of high quality media for preservation purposes (gold CDs perhaps) may reduce the need for refreshing, and help diminish losses from media deterioration, but that is all it does. It has no impact on any other potential source of loss (Kenney et al., 2003). We need to keep in mind that relying on strategies and practices that focus on media storage have the potential for endangering content by providing a false sense of security (Kenney et al., 2003). We should also keep in mind that there is still a considerable lack of awareness about the real issues of digital preservation. An indication of this lack is the BBC Domesday Project, where IT industry vendors offered solutions based on extending the life of CD-ROMs for 100 years. As noted in Chapter 2, the myth that long-lived media equals long-lived preservation is still worryingly popular (Abbott, 2003, p.10). Despite the likelihood that even modest improvements which produce storage media with larger per unit storage capacities and greater tolerance to variations in temperature and humidity will lower preservation (Hedstrom, 1998, p.197), it is doubtful that there will be sufcient benet to warrant serious attention. There is, perhaps, an analogy with the efforts made by the library and publishing industries, archivists, and authors to inuence the manufacture of paper so that it was less acidic and longer lasting (Harvey, 1993, pp.192194) but the benets in that case were signicant. For digital materials, the outcome of this way of thinking (which is perhaps pre-digital preservation paradigm thinking) is considerably less productive.
Analogue backups
Making analogue backups is a non-solution in the sense that it preserves only part of the digital materials and loses what Rothenberg refers to as their core digital attributes, including their machine-readability, interactive abilities, and other aspects of their functionality (Rothenberg, 1999a, p.9). This approach is most commonly discussed in terms of making copies of the textual and image information content of the digital material by printing them to paper or transferring them to microlm. Early attempts to preserve digital materials often focused on copying them from relatively unstable digital storage media to formats known through extensive experience to be stable, such as paper and microlm. Hedstrom remarked in 1998 that it was probably the most commonly used preservation strategy because, lacking more robust and cost-effective migration strategies at that time, copying to more stable formats, even if they were analogue, offered advantages as a method of last resort (Hedstrom, 1998, p.194). Although it has largely been discounted as a viable preservation approach, it may on occasion still be useful. An Australian digital preservation specialist interviewed in 2004 commented:
when we rst started thinking about preserving web sites, and capturing them was at that stage seen to be too difcult, I considered for a time just at least printing out all the web pages that we could identify . . . on good colour printers and storing them as representations of at least what was on the Web in 1996 . . . [but] I thought that was inappropriate as a form of preservation . . . because it had to be digital. Well in fact now I really rue that decision because at least we would have had at least some record of what was happening in 1996 on the Web. More specically, making analogue backups or copies is dened as converting certain valuable digital resources to a stable analogue medium (Jones and Beagrie, 2001, p.109), thus shifting the preservation burden to an analogue copy in place of the digital object (UNESCO, 2003, p.146). The stable analogue media suggested are paper (usually permanent paper), microlm (silver halide or produced to preservation standards), and HD Rosetta, explained as nickel plates onto which document images are micro-engraved using an ion beam by HDRosetta technology; viewing with optical magniers is required (UNESCO, 2003, p.146; Norsam Technologies, 2001). Making analogue backups is useful as a preservation practice for a limited number of categories of materials. For this material, it preserves the information content in a form free from the threat of technological obsolescence because it either does not need equipment to access it (as in paper) or requires only relatively simple non-digital equipment (such as a microlm reader). These analogue formats are likely to remain accessible for some hundreds of years if they are produced to preservation standards (such as the standards dened for permanent paper or archival-quality microlm) and stored in appropriate conditions. This approach provides a simpler preservation alternative, as analogue materials may be preserved for the long-term using traditional preservation methods (UNESCO, 2003, p.146). However, the advantages of making analogue copies are countered by some signicant limitations. Jones and Beagrie consider that making analogue copies is not a digital preservation strategy because it does not preserve the digital characteristics of the materials it is applied to, but is, rather, a way of preserving the information content of digital materials and is only applicable to small-scale operations (Jones and Beagrie, 2001, p.103). The major limitation is that analogue copies do not possess the digital qualities of the original materials. For instance, there is a loss of functionality, such as the ability to carry out calculations in spreadsheets or to search the digital data, which may be an issue for some digital materials but not for others (UNESCO, 2003, pp.146147). The ability to share digital materials readily is reduced, as is the potential for lossless transferability (Kenney et al., 2003) and storage efciency. Making analogue copies is best considered as a pragmatic interim strategy applicable only to a limited category of digital materials for which loss of functionality is less important, pending the development of more appropriate digital preservation strategies (Jones and Beagrie, 2001, p.109). The UNESCO Guidelines suggest that it may be required as a last resort where no other strategy is available and such limited accessibility is better than no accessibility at all (UNESCO, 2003, p.147). It is likely to be used only where the full functionality of the digital materials being copied is not considered to be important to preserve, such as textual, image or data type documents that require no functionality above at display (UNESCO, 2003, p.147).
Non-solutions
125
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Two examples are provided in the UNESCO Guidelines. The rst is printing out to paper of a database in an obsolete proprietary format which cannot be extracted or migrated; the database is retained in its digital version to allow for the possibility that it can be accessed at a later date. The second example is the creation of COM (Computer Output Microlm) of digital image masters created during a digitization programme as a backup for preservation purposes (UNESCO, 2003, p.146). If analogue copies are made for preservation purposes, two conditions should be adhered to. The analogue formats should be produced to archival standards and housed in archival conditions. The original digital objects should, where possible, be retained to allow for the possibility of access mechanisms becoming available in the future (UNESCO, 2003, p.147). It is possible that storing analogue formats in archival conditions over long periods of time may not be feasible on technical and/or nancial grounds, so these factors should be considered. To summarize: Text and monochromatic still images are the most amenable to this kind of transfer. Given the cost and limitations of analog backups, and their relevance to only certain classes of documents, the technique only makes sense for documents whose contents merit the highest level of redundancy and protection from loss (Kenney et al., 2003).
Policy development
Developing a digital preservation policy is suggested by Howell as a prerequisite to digital preservation (Howell, 2001, p.142). This is best thought of as a general principle relating to most, if not all, aspects of managing an information service, and is not addressed further here.
Standards
Standards in the context of non-solutions relates to a principle, rather than a strategy or practice, if we apply our Chapter 6 denitions. Attention to the use of standards provides a more favourable environment in which digital preservation can be successful: their widespread use can be considered a prerequisite for effective preservation. The increase in standards development for digital preservation indicates, it is suggested, a certain maturation of the eld (Hodge and Frangakis, 2004, p.48), and is of considerable importance because it promotes collaboration and interoperability among organizations. Adelstein describes the current state of ISO standards development for preservation of digital materials, which focus on media quality, storage, and handling (Adelstein, 2003). The use of standard data formats is not, however, a non-solution and is noted in Chapter 8. One principle that lies behind the promotion of standards that are stable and have been widely adopted is that there is an increased likelihood that they will be supported and will remain viable over a longer period. This is not always the case with proprietary standards or standards designed for use only with specic software and/or hardware. If standards are open source (that is nonproprietary), so much the better, for this reduces the risk of obsolescence if there is a large user base willing to ensure that the standards are maintained. Jones and Beagrie suggest that the widespread adoption of open standards thus can defer inaccessibility of digital resources due to technological obsolescence (Jones
and Beagrie, 2001, p.107). Other advantages that are perceived to accrue from standardization are that costs will be reduced because there are fewer and less complex variations to deal with: for example, economies of scale could be achieved when migrating data (Jones and Beagrie, 2001, p.107). The UNESCO Guidelines add that the use for widely available standards is more likely to allow re-interpretation of the data or re-construction of tools in the future, if necessary but notes that use of standards is an investment strategy, that is, it involves investment of effort at the start of the process (UNESCO, 2003, pp.127128) so there are resourcing implications. Rothenberg sounds a word of caution: standards are not enough and are not realistic in the foreseeable future because using successive, evolving standards would require translation, but translation between standards is rarely reversible without loss, so this cannot reconstruct an original artifact (Rothenberg, 2003).
Digital archaeology
Digital archaeology, often referred to as data recovery, denotes a set of techniques that are applied as a last resort, and is therefore not a solution in any sense. The term applies to methods and procedures to rescue content from damaged media or from obsolete or damaged hardware and software environments and involves specialized techniques to recover bitstreams from media that has been rendered unreadable, either due to physical damage or hardware failure and is explicitly an emergency recovery strategy (Kenney et al., 2003). Special facilities, equipment and expertise are required and it is usually carried out by specialized data recovery companies whose expertise has established that it is possible to recover data from a wide range of media types. The recovery of the data does not necessarily lead to recovery of the ability to understand that data, although it is a necessary pre-condition for it (see Dunning, 2001 for an example). The UNESCO Guidelines caution us against relying on digital archaeology and remind us that it is a very unreliable and high-risk substitute for a current and active preservation programme (UNESCO, 2003, p.148). For one thing, it is very expensive and most materials would not justify the costs of recovering them. But more important, it is not reliable; there is no guarantee that digital materials can be recovered, and, even if the data is recovered, there is no certainty that it will be intelligible (Jones and Beagrie, 2001, p.110; UNESCO, 2003, pp.148149). Ross and Gows 1999 study Digital Archaeology: Rescuing Neglected and Damaged Data Resources suggests that data recovery should be unnecessary if good disaster planning is in place, but this very rarely is the case. They note that (and this is the sting in the tail) with sufcient resources much material that most of us would expect to be lost can be recovered (Ross and Gow, 1999, p.iii). Their detailed report is essential reading for further information about digital archaeology and its techniques.
Preserve technology approaches

Approaches located at the preserve technology end of the digital preservation spectrum are those that attempt to keep data in specic logical or physical formats and use technology originally associated with those formats to access
127
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
the data and reproduce the objects (Thibodeau, 2002). That is, there is as little change as possible, preferably none, to both the digital materials (which remain exactly as received) and the software, operating system, and hardware that these materials were originally developed to be rendered on. Strategies and practices at the extremes of the preserve technology end of the spectrum allow ideally no change at all. A little further along the spectrum is emulation of old technology, where change is allowed but the original conditions are maintained as far as possible. Of the two approaches noted in this preserve technology category, one emulation is commonly considered as viable for long-term preservation, but the other maintain original technology is not. The UNESCO Guidelines take the view that technology preservation is a short-term strategy (likely to work best over the short-term only) and classies emulation as medium- to longterm, likely to work over longer periods (UNESCO, 2003, p.126).
Technology preservation
The preserve technology approach to digital preservation is also referred to by other terms which, taken together, dene its scope: preserve technology; maintain old technology; maintain museums of working computing equipment, software and documentation; and the related strategy technology watch. This approach, which exudes a certain technological bravado (Rothenberg, 1999a, p.12), requires that obsolete hardware and software are maintained in working condition so that the digital materials retained by an institution can be accessed. It has a long history, being perhaps the most immediately obvious approach to take when obsolescence is imminent, but our experience with this approach has shown us that it is an expensive ultimately too expensive approach. This experience has also indicated that, given the large number of formats, hardware and software that become obsolete (a number that is increasing over time), it is not a viable option for anything but the short term. It is also an approach that most, if not all, institutions that undertake digital preservation have applied, and has an important role as a rst line of defence: it is the most basic, and in some ways the most important rst step in preserving access if no other strategy is in place. If the hardware and software required for access are discarded before other strategies are available, it may be effectively impossible to provide later access without expensive and uncertain data recovery work (UNESCO, 2003, p.135). One example of technology preservation is maintaining old disk drives, as in the State Library of New South Wales where a 512 inch oppy disk drive was maintained so that material on these disks that had been lodged under legal deposit legislation could be copied to a more current medium (Howell, 2001, p.142). Technology preservation is considered as useful only in the short term, ultimately a dead end because obsolete technology cannot be maintained in a functional condition indenitely. At best it can extend the window of access for obsolete media and le formats but if it is attempted it is heavily resourceintensive (Kenney et al., 2003). The experience of preserving technology in audiovisual archives over many decades provides clear evidence of this. Edmondson (2004) provides a starting point for further investigation on this point.
Technology preservation is perceived to have many positives. Authenticity is preserved, both of the digital materials (which are unaltered), and of the technology and software platforms that render them (which are also not changed). Thus the functionality, look and feel of the original (Jones and Beagrie, 2001, p.106) are retained. Preserving the technology provides the benet of buying time, by delaying the need to apply practices that will ensure long-term preservation. An additional benet may be that the documentation for the software and hardware preserved is likely to be of future use when implementing other strategies (UNESCO, 2003, pp.135136). But there are many less positive factors in technology preservation. It becomes more and more expensive as replacement parts for hardware and the expertise to maintain it become scarce: technical support will inevitably disappear within a relatively short timeframe (Jones and Beagrie, 2001, p.106). This effect may, however, be ameliorated by sharing expertise, equipment and parts among a number of organizations (UNESCO, 2003, p.136). The access it allows to obsolescent and obsolete digital materials is extended, but only for a short period, according to the UNESCO Guidelines only for a period as narrow as ve to ten years from the time the original format is superseded (UNESCO, 2003, p.135). To be effective this approach requires access to documentation, such as manuals, for the software and hardware, which also needs to be procured and managed. To work effectively, this preserve technology approach has specic requirements, which are articulated in the UNESCO Guidelines. It requires that the hardware and software that will be needed to provide access is actively identied and its move towards obsolescence monitored (This is discussed in the Technology watch section below.) Active and continuing arrangements to maintain hardware and to license software should be put in place. The expertise needed to maintain equipment and software should ideally not reside in one person but should be shared (UNESCO, 2003, p.136). It is worth reiterating that technology preservation is useful in the toolkit of approaches and practices for digital preservation, but, in doing so, it must also be noted that it can be no more than an interim strategy because of its limitations. The UNESCO Guidelines go further by suggesting that it should be a matter of principle that if the required access software is available, it should be sought and retained at least until another strategy has been put in place. Technology preservation is best considered as an initial strategy for all preservation programmes, in the absence of longer term strategies or while they are being developed and may have specic application to maintaining software that is used to support other strategies (UNESCO, 2003, p.136). Museums of computing have been established, such as the Computer History Museum (www.computerhistory.org), but these are established principally for museum purposes and maintaining working computers is unlikely to be their primary intention. Some envisage that a widespread industry will develop, the equivalent of Kinkos where theyll have every ancient computer available, whose customers will drop in to read and copy their old data les (Peter Schwartz quoted in Hafner, 2004).
Technology watch
A strategy related to technology preservation is technology watch. Because the process of technology obsolescence is rapid, and storage media and le formats
129
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
change equally rapidly, there is a need to monitor the rate of change so that access to digital materials in danger of being lost because of this obsolescence can be maintained. This process of technology watch acts as a trigger to action, by ensuring that equipment is maintained until another preservation practice is applied to at-risk material, or by migrating to a more stable storage medium (Bennett, 1997, p.6). The process is simply described: 1) identify at risk technology in a collection 2) monitor its rate of obsolescence at regular intervals 3) take action when a technology is in danger of no longer being supported. At the international level, Bennett proposes a survey of organizations to develop base-line data, updated regularly by consultations with a group of Wise Practitioners to forecast technology trends (Bennett, 1997, p.27). At the institutional level, a rst step is likely to be a survey of digital materials present in a collection to identify the formats, media, and software requirements present. A risk assessment approach from which a plan of action is developed may be useful (Jones and Beagrie, 2001, pp.132,134). This initial identication process is assisted if the storage media and software and hardware requirements are noted in the catalogue records of a collection, but this is not always the case. When a survey of digital materials in the collections of the State Library of Victoria was carried out in 2003, it was initially assumed that the Librarys catalogue could be used to identify relevant material. In the event, the initial simple searches of the catalogue identied only about 3000 items from a total of about 12 000 items, which were located by carrying out complex searches and shelf checks. Learning from this experience, and to track the size and the nature of its growing digital collection, the State Library of Victoria now puts a digital ag into a specied eld of its catalogue records. The National Library of Australia developed a list of hardware and software as a basis for agging changes in technology (Woodyard, 1999). Identication will be considerably assisted in the future if the metadata records for all current additions to the collection contain this data, so that identifying material at risk from technology obsolescence is straightforward. A list of hardware and software available in an organization should also be maintained and used to plan for their replacement (Jones and Beagrie, 2001, p.134). The Digital Preservation Coalition publishes a series of Technology Watch Reports, available from its web site (www.dpconline.org/graphics/reports). An example is its report on mass archival storage systems, based on experience at the British Library, published in February 2005.
Emulation
The battle lines were drawn in the 1990s between migration and emulation as the preservation strategy most likely to succeed. In the event neither has dominated, as we learn to place less trust in a single-strategy salvation and to develop ways of working and thinking that accommodate several approaches simultaneously. Emulation is based on the principle of simplify[ing] digital preservation by eliminating the need to keep old hardware working (Thibodeau, 2002, p.19). Emulation and emulators are variously described and dened. In terms of preservation, emulation is a means of overcoming technological obsolescence
of hardware and software by developing techniques for imitating obsolete systems on future generations of computers (Jones and Beagrie, 2001, p.105). Emulation is the ability of a program or device to imitate another program or device, using software that makes one technology behave as another (UNESCO, 2003, p.144). It combines software and hardware to reproduce in all essential characteristics the performance of another computer of a different design, allowing programs or media designed for a particular environment to operate in a different, usually newer environment. Emulators, programs that translate code and instructions from one computing environment so it can be properly executed in another, are required (Kenney et al., 2003). Emulation is a well-established principle in the computing industry. Emultion can be of operating systems (for example, VirtualPC, which allows PC software to be run on an Apple Macintosh), of hardware platforms (for example, Kaypro or Apple II machines can be emulated on a PC), and of software applications (for example, arcade games emulated through the MAME project). One emulator in common use is built into the Apple Macintosh operating system and enables this computer to run software developed for earlier Apple computers. Printers are often designed to emulate Hewlett-Packard printers: here, notes the Webopedia denition (www.webopedia.com), emulation tricks the device into believing it [is] really some other device. Terminal emulation is also common, so that a PC, for example, can be used as a terminal connected to a larger computer; these were once very common in mainframe computer environments. Although emulation is often associated with computer games, such as emulators that allow Sony PlayStation games to be played on a PC, it is an essential part of computing in all areas. The web is a fruitful source of information about emulators (e.g. www.directory.net/Computers/Emulators). Both hardware emulation and software emulation have been experimented with to determine their feasibility for preserving digital materials. The aim is to make future technologies behave like the original environment of a preserved digital object, so that the original object could be presented in its original form from the original data stream (UNESCO, 2003, p.144). Emulation of hardware has the attraction of being applicable to a wide range and large amounts of digital materials, because it would allow a range of systems and digital objects to operate, thus solving the problem for a very wide range of digital objects (UNESCO, 2003, p.144). The same widespread applicability applies to emulation of operating systems. Emulation of software applications is less in favour, because it is more limited in its use and the effort and level of skill required to develop a complex piece of emulator software that can be used for only a small number of digital materials may be too high for this to be an option in most preservation situations. Emulation has been investigated in several major projects, most notably, the CAMiLEON (Creative Archiving at Michigan and Leeds) project which investigated emulation, testing available emulators and constructing an emulator for the BBC Domesday Project. Among its conclusions was that emulation is not necessarily superior to migration for preserving the original look and feel of complex digital objects. (The projects web site (www.si.umich.edu/ CAMILEON) provides more information.) However, more research was needed, as this was a study of limited scope (Hedstrom and Lampe, 2001). Project NEDLIB (Networked European Deposit Library) ran from 1998 to 2000 and was based at the Koninklijke Bibliotheek in The Hague. One of its activities was to
131
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
conduct an experiment using commercial emulation tools to investigate the feasibility of emulation for digital preservation. This experiment, conducted by Rothenberg, was limited in its scope. It concluded that emulation should work in principle, but that further investigation is required to demonstrate that it can also work in practice (Rothenberg, 2000). The projects web site (www.kb.nl/ coop/nedlib) gives more information. Jeff Rothenbergs name is rmly linked with emulation as a digital preservation strategy. He has been a strong proponent of emulation as the only digital strategy that is likely to be effective. His argument is that the only way to make sure that digital materials are preserved in their native form (so that they are experienced appropriately by future users) and are not corrupted is to run software that interprets the bit-stream correctly. Because software can behave differently in different contexts, we should not expect software developed in the future to reproduce the behaviour of obsolete software. To ensure the appropriate behaviour of digital materials it is necessary to run the original software used to view the document in its native form. Rothenberg points out that all digital materials depend on software, and in particular many new kinds of digital materials are inherently digital and cannot be meaningfully represented as page images (Rothenberg, 2003). This means that preservation strategies such as saving page images are of little use for the preservation of much digital material. Emulation, according to Rothenberg, is the only preservation approach that has multiple advantages and capabilities, among them preserving executable digital objects (objects in which software programs are embedded), providing a single, consistent way of preserving all kinds of digital materials, reducing the effort expended in preserving individual artifacts (except for copying the bit-stream onto new media), and minimizing the need to understand record formats. Despite his strong advocacy of emulation, Rothenberg suggests that a mixed strategy approach is most feasible, with emulation used if original behavior is needed; or as a cheap backup, to preserve everything (Rothenberg, 2003). Other authorities also suggest that emulation has potential. Jones and Beagrie consider it to be one of only two primary strategies (those suitable for medium to long-term preservation of digital materials), the other being migration (Jones and Beagrie, 2001, p.103). They suggest that it has the advantage over other methods of recreating the look and feel and functionality of the original digital material, as well as the potential for avoiding the high costs associated with repeated migration. Emulation, by their judgement, has good prospects for preserving complex digital materials (Jones and Beagrie, 2001, p.105). The UNESCO Guidelines note that emulation is already well established and understood in computing, that many emulators already exist for a variety of hardware and software platforms, and that it has the potential to allow a range of digital objects to be recreated with full functionality, including software objects, using the original, untransformed data stream in combination with original preserved software (UNESCO, 2003, p.145). Not all share Rothenbergs enthusiasm, among them Granger (2000). The arguments against emulation form a list as long as the arguments in its favour (Jones and Beagrie, 2001, p.105; UNESCO, 2003, p.145). Chief among them is that emulation has not been sufciently tested in practice. Also high on the list of disadvantages is the high cost of developing emulators, which may be greater than the costs of repeated migration, because it requires high levels of
expertise to write complex software. There is some scepticism about the ability of emulation to do all that is claimed, and it may not be possible to emulate fully all of the functionality of the original, nor all of its look and feel. The lack of adequate documentation of hardware and software may frustrate emulation attempts. Copyright issues associated with ownership of software code may impede emulator development. Users may have problems in interacting with archaic applications operating under emulation, and the need either to migrate the emulators themselves or emulate the emulators raises the interesting spectre of layers upon layers of emulators (UNESCO, 2003, p.145). If emulation is to be attempted, then the requirements are many and varied. Appropriate expertise is, of course, essential. Documentation of the systems to be emulated needs to be comprehensive and accurate. The emulation software should be written in open source code, using a standard programming language with good prospects for longevity and future compatibility (UNESCO, 2003, p.145) and following best industry practice, including thorough documentation. Probably the most widely reported emulation project carried out for preservation purposes to date is the BBC Domesday Project. Abbott (2003), Darlington, Finney and Pearce (2003), Mellor (2003), National Archives (U.K.) (2004?) are just some of the reports on this project; the CAMiLEON web site (www.si.umich. edu/CAMILEON/domesday/domesday.html) is also a useful source. The original Domesday Project, undertaken from 1984 to 1986, surveyed the UK to celebrate the 900th birthday of the Domesday Book. It cost about ?2.5 million and involved about one million school children from 14 000 British schools. The resulting images and text were recorded on two 12-inch videodisks that were accessed using a LV-ROM (LaserVision Read Only memory) player attached to a BBC Master computer with additional software and hardware (Abbott, 2003, p.7). As part of its activities the CAMiLEON Project developed an emulation of the original Domesday system hardware. The process of developing this emulation involved migrating the data les from the videodisks to current media, and developing software that emulates the BBC Master computer and the laserdisk player (Mellor, 2003). Image les were re-digitized from the original one-inch analogue videotapes (Darlington, Finney and Pearce, 2003). The need to avoid the obsolescence of the emulation software was kept in mind: ideally, the software should not be limited to operation on any specic operating system or type of computer so that it will be easier to run on future computers. In the end, however, the BBC Domesday emulator that was developed runs only on the Windows operating system (Mellor, 2003, p.8). Signicant lessons were learned from the BBC Domesday Project. This emulation project was hampered by lack of documentation and software to test the emulation, but was fortunate to have available a working original system, albeit a fragile one. This was important because it allowed the developers to compare with and validate the migrated system, which has special signicance in a multimedia system where the look-and-feel and user interaction is important (Darlington, Finney and Pearce, 2003). Wheatley, one of the team who worked on this project, summarizes the issues: Most of the really difcult problems we faced were due to the long time gap between the creation of Domesday and its preservation. If we had conducted the rescue 10 years earlier it would have been far easier. The
133
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
timeliness of preservation work is a crucial issue that Domesday really underlines. Would we be able to rescue Domesday if we left it another 10 years? Im sure we could, but it would be at far greater expense (Abbott, 2003, p.10). Emulation is likely to play a signicant role as a major preservation strategy. It has been sufciently well tested to show promise. What is now required to exploit that promise is the allocation of signicant resources, probably through collaborative action, to develop a range of emulators. Emulation is unlikely to be the magic bullet (Lynch, 2004), the single solution to digital preservation, as it is sometimes promoted. Nor is it likely to replace migration as a primary digital preservation strategy, because emulators will themselves need to be migrated. Holdsworth and Wheatley remind us that emulation should not be over-sold as the answer to all digital preservation issues. It is just part of the armoury necessary for defending our digital heritage against the ravages of time in a world where innovation (and hence change) is highly prized (Holdsworth and Wheatley, 2001). We can envisage that emulation will be used increasingly as more emulators are developed, in specic situations, such as for complex digital materials or for those which contain executable software that will only run on specic hardware, or for digital materials that need to be viewed in their original environment. These, of course, presuppose that suitable emulators are already available or that the expertise is available to develop them (UNESCO, 2003, p.145).
The Universal Virtual Computer

An extension of emulation is the Universal Virtual Computer. This program is constructed independently of any existing hardware or software, so that it is independent, too, of time. It would simulate the basic architecture that every computer has had since the beginning: memory, a sequence of registers, and rules for how to move information among them. Computer users could create and save digital les using the application software of their choice; when a le was saved, though, it would also be backed up in a way that could be read by the universal computer (Tristram, 2002, quoting Raymond Lorie). The Universal Virtual Computer is general and basic, and its simplicity ensures that it will be relatively easy to write an emulator for [it] in the future on the real machine being used at that time (Lorie, 2002, p.1). The Universal Virtual Computer has been developed by the Koninklijke Bibliotheek and IBM as a proof-of-concept prototype that successfully demonstrated that the concepts are quite sound, can be applied to the archiving of PDF documents, and can be implemented (Lorie, 2002, p.2). More recently the Koninklijke Bibliotheek and IBM have developed and implemented a UVC tool for images in JPEG format (Van Wijngaarden and Oltmans, 2004). The signicance of such a virtual computer is that it provides a single platform, with the potential for minimizing the amount of effort required to handle diverse combinations of hardware and software. However, the Universal Virtual Computer approach is still experimental. If it becomes more widespread, it represents a strategy that requires
considerable investment at the beginning of the preservation process (UNESCO, 2003, pp.133134).
Conclusion
This chapter has described a number of strategies and practices currently applied or being tested for digital preservation. They are all focused on the principle that alteration of digital materials must be kept to the minimum and that the technology (hardware and software) to access these materials is kept operational or is emulated. Of the range of strategies and practices noted here, some are interim measures and only one emulation is generally considered to be viable for the long-term preservation of digital materials. The next chapter considers a range of approaches from the other end of the preservation spectrum: preserve objects.
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Chapter 8
Preserve Objects Approaches: New Frontiers?
Introduction
Migration has been the only serious candidate thus far for preservation of large scale archives (Granger, 2000) Migration is essentially an approach based on wishful thinking (Rothenberg, 1999a, p.15) Handling le formats is an essential part of a long-term archive . . . most le formats fall out of fashion within a few decades, and unless action is taken at an early stage, many archived les will be incomprehensible blocks of bits (Clausen, 2004). Let us return briey to the beginning of the previous chapter and its derivation, from the work of Thibodeau and Rothenberg, of three categories of preservation strategies and practices: 1) Non-solutions 2) Preserve Technology approaches 3) Preserve Objects approaches. This chapter looks into the third of these groupings, the Preserve Objects approaches which are at the opposite end of Thibodeaus spectrum (see Figure 6.3) from Preserve Technology approaches. Like Chapter 7, it relies heavily on two key sources: Preservation Management of Digital Materials: A Handbook by Maggie Jones and Neil Beagrie (Jones and Beagrie, 2001), and the UNESCO Guidelines for the Preservation of Digital Heritage (UNESCO, 2003). The Appendix contains several case studies relevant to this chapter. Case Study 1 examines digital storage at the National Film and Sound Archive and illustrates the topic of migration. Case Study 3, which notes digital preservation activities at the
136 Preserve Objects Approaches: New Frontiers?
National Archives of Australia, illustrates the topics of normalizing and XML. VERS (Victorian Electronic Records Strategy) is the subject of Case Study 4, and illustrates the topic of encapsulation. Case Study 5 describes migration at the Australian government Department of Family and Community Services.
Preserve Objects approaches

The preserve objects approach attempts to preserve the essential characteristics of digital materials without reliance on specic hardware or software. That is, the digital materials may be altered, although it is preferable if they are not, but only to the extent that is required to make them able to be rendered by current and future technology (hardware and software), without compromising the essence or essential elements (discussed in Chapter 5) of that material. These essential elements are dened explicitly and independently of specic hardware or software (Thibodeau, 2002). The preserve objects approach can be contrasted with the other end of the spectrum, the preserve technology approach, where, as stated in Chapter 7, as little change as possible preferably none occurs to both the digital materials and the software, operating system, and hardware that these materials were originally developed to be rendered on. The strategies and techniques encompassed by the preserve technology approach attempt to keep data in specic logical or physical formats and to use technology originally associated with those formats to access the data and reproduce the objects (Thibodeau, 2002). There are other positions along the spectrum between these two poles; for example, in the middle are methods that migrate data formats as technology changes, enabling use of state-of-theart technology for discovery, access, and reproduction (Thibodeau, 2002). This chapter is concerned with the range of strategies and practices that seek to accommodate changes in technology without compromising the digital materials themselves or the ability to render their essential elements accurately and meaningfully. These strategies and practices are grouped for the purposes of this chapter as: those based on copying the bit-stream (backup and restore, bit-stream copying, refreshing, replication, mass digital storage systems) those centered around data formats (standard data formats, Universal Preservation Format) migration, which requires the bit-stream to be copied, and for which significant knowledge about data formats is required (including migration on demand and the use of viewers) encapsulation, which is based on packaging the digital object with metadata and other associated information. Only one of these groups of strategies and practices migration has been implemented to any great extent. Jones and Beagrie consider migration to be a primary strategy: one which might be selected . . . for medium to long-term preservation of digital materials (Jones and Beagrie, 2001, p.102). Other strategies described in this chapter, such as standardizing data formats and encapsulation, are secondary strategies those which might be employed in the short to medium term either by the repository with long-term preservation responsibility and/or by those with a more transient interest in the materials
Bit-stream copying, refreshing, and replication 137
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
(Jones and Beagrie, 2001, p.102) which may precede, substantially defer the need for, or strengthen primary strategies. Some strategies noted in this chapter are investment strategies, in terms of the UNESCO Guidelines classication of digital preservation strategies (UNESCO, 2003, pp.126149); they primarily involve investment of effort at the start. Encapsulation and restricting the range of formats to be managed fall into this category. Other strategies noted (backwards compatibility and version migration, and migration) in this chapter are short-term strategies, those likely to work best over the short-term only. One (the use of viewers) can be characterized as a medium- to long-term strategy likely to work over longer periods. Two of the strategies and practices included in this chapter are noted as current front-runners as long-term strategies and that have been shown to work over short periods of time (UNESCO, 2003, pp.124126). These are using standards, especially standards for data structuring, and migration. These strategies and practices in general support the principles suggested in the UNESCO Guidelines.
Bit-stream copying, refreshing, and replication

Some denitions are called for. Bit-stream copying, sometimes referred to as backup and restore, is the process of making an exact duplicate of a digital object. Refreshing is the copying of data onto new media. Replication, also known as redundancy, refers to keeping multiple copies. All of these processes involve making copies, but each differs signicantly from the other.
Bit-stream copying
Bit-stream copying is the process of making an exact duplicate of a digital object (Kenney et al., 2003). Procedures for bit-stream copying are well understood. They are commonly practised as backup and restore, that is, backing up (copying) computer les on a regular basis and restoring them if the data in the primary source (the les from which the backup copy or copies are made) are corrupted or destroyed. Regular backup routines are a key component of the operation of any computing facility. In making copies of the bit-stream, well-established techniques that ensure that the data is copied accurately are applied, such as check-sums and digests (Howell, 2001, p.144). The backup les are often stored at sites away from the main site of operation, which offers additional security in the event that one site is affected by disaster. This is usually referred to as remote storage. Although it is necessary for all digital preservation strategies, bit-stream copying is not a long-term strategy because it does not address the key factors that cause digital deterioration, mainly obsolescence of hardware and software. It is effective as a risk management technique to minimize the possibility of data loss through failure of hardware, media deterioration, sabotage, natural disasters or other events. It should be considered as the minimum maintenance strategy for even the most lightly valued, ephemeral data (Kenney et al., 2003).
Refreshing
Refreshing is a renement of bit-stream copying. Whereas bit-stream copying is carried out so that a backup copy exists in case there is a problem with the
primary source les, refreshing takes a longer-term view. It is typically carried out in order to copy data from one long-term storage medium to another of the same type (Kenney et al., 2003), for example, from a DAT tape that is becoming unstable to a new DAT tape. The bit-stream is not altered. As with bit-stream copying, refreshing is a necessary component of any successful digital preservation program but does not address issues of obsolescence (Kenney et al., 2003).
Replication
Keeping multiple copies replication is a long-established preservation technique in libraries, valued for its built-in high redundancy and protections against loss of information through vandalism, theft and disaster (Howell, 2001, p.143). For example, a copy of a book may be kept in secure environment-controlled storage as a preservation copy, with other copies available for use. Kenney et al. (2003) distinguish among several uses of the term replication: as bit-stream copying, as refreshing, and, in the sense that it is used here, as a consortial form of copying of digital materials. This is demonstrated in the LOCKSS project, where the intention is to enhance the longevity of digital documents while maintaining their authenticity and integrity through copying and the use of multiple storage locations (Kenney et al., 2003). The progenitors of LOCKSS, as reported by Howell (2001, p.143), describe replication as a fault-tolerant paradigm which works by acquiring copies, distributing them around the world so that some can readily be located but others are hard to nd, and lending or replicating copies when others need access. Replication of digital materials had been noted as a feasible preservation strategy before it was implemented in the LOCKSS project. Exon suggested that mirror sites around the world would store selected digital information according to international agreement, so that when the tides of history turn against previously stable and peaceful areas of the world, their history can be saved elsewhere, just like the Elgin marbles (Exon, 1995). LOCKSS was developed for the preservation of e-journals. It is based on open source software that harvests, stores, and copies digital content using standard desktop computers and ensures accuracy of the digital material through peerto-peer polling. It is inexpensive because it does not require costly hardware, the software is free, and, as claimed on the LOCKSS web site, very little technical administration is required. In 2004 more than 80 libraries and 50 publishers were using the LOCKSS software, and further development was taking place through collaborative research to expand its capabilities. The LOCKSS web site (lockss.stanford.edu) provides a detailed summary of how the system works. An inexpensive personal computer running LOCKSS software collects specied e-journal content, for which the publishers permission to collect has been secured, using a web crawler. It compares continually the content collected with the same content that has been collected by others in the network and ensures that the content is identical. It also allows access to this content by authorized users and provides administrative functions. In order to allow the LOCKSS crawler software access to their content, publishers need to give permission to the LOCKSS system. The basis of the preservation function of LOCKSS is the continual checking of digital content against other copies and the repairing of discrepancies identied by comparing copies through polling. The LOCKSS web site provides more details:
Standard data formats 139
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
At intervals appliances [the name given to the personal computers with LOCKSS software] take part in polls, voting on the digest of some part of the content they have in common. If the content in one appliance is damaged or incomplete that appliance will lose the poll, and it can repair the content from other appliances. This cooperation between the appliances avoids the need to back them up individually. It also provides unambiguous reassurance that the system is performing its function and that the correct content will be available to readers when they try to access it. The more organizations that preserve given content, the stronger the guarantee they each get of continued access. LOCKSS is also promoting a range of related projects to investigate preservation of other kinds of e-publications, for example humanities e-journals, government documents, newspapers, electronic theses and dissertations, web sites, and materials in special collections. In November 2004 the LOCKSS project announced that it has successfully demonstrated transparent format migration of preserved web content (Rosenthal et al., 2005). Other information is readily available to supplement the operational and technical detail given on the LOCKSS web site, such as an early paper describing the concept and reporting on its test phase (Reich and Rosenthal, 2000) and a fuller description of production-scale activities (Reich, 2002).
Standard data formats

Understanding le formats is a necessary precondition for the effective use of many digital preservation techniques. In order to access and display digital content it is necessary to decipher the bit-stream, to learn what the information in that bit-stream represents. This is the role that le formats play. They dene how to interpret the bits in the bit-stream. File formats are a crucial layer, indeed a hinge between the bits in storage and their meaningful interpretation . . . one of the core issues of any digital preservation approach, and le format obsolescence is a major challenge for anybody wanting to preserve digital les (ERPATraining, 2004, p.1). File formats would not pose so great a challenge to digital preservation if there were fewer of them and they were more accessible. The wide range of standards and formats that need to be handled is illustrated in the following examples. A survey of the National Library of Australias collections in 1996 identied 134 CDs and 721 oppy disks, of which 172 were 512-inch and 549 were 312-inch disks. As well as PC and Macintosh, other operating systems were required to access these disks: Apple II (various varieties), BBC, Commodore 64 or 128, Apricot PC, Microbee 64, and unknown. A signicant number of the disks were over ve years old (Woodyard and Webb, 1997). Analysis of the data on a sample of 64 disks, chosen to represent variables such as operating system, software, disk size, and age, identied many problems, including lack of appropriate hardware or software, disk errors, and illegal characters in le names. Of the 64, 37 per cent were not accessible at all. Of the 40 disks that could be copied, 55 per cent produced functional les after copying, 12 per cent were not
functional after copying and 33 per cent were probably largely functional but the National Library of Australia did not have the software needed to test this (Woodyard, 1998). In a survey of RLG member institutions in the same year, at least 24 different storage formats were identied. Of the 36 institutions with digital holdings, 24 maintained these in six or more different formats, with 10 or more different formats in 13 institutions. The most common le formats were image les, text les with mark-up, and ASCII les, with word-processing les, audio, video, and spreadsheets also well represented (Hedstrom and Montgomery, 1999, pp.910). Another example illustrating the range of standards and le formats that need to be handled in everyday use comes from a proposal for a DOMS (Digital Objects Management System) at an Australian university:
Open system/ Internet standards DOI HTTP/HTML Z39.50 Java XML Resource Description Framework (RDF) MIME
Metadata standards MARC21 Dublin Core EAD IMS SCORM ONIX
File formats
Image formats: PNG, PSD, TIFF, GIF, JPEG, BMP, and other common graphic le types Audio formats: AU, AIFF, WAV, RealAudio, Audio CD and other common audio le types Video formats: MPEG, QuickTime, RealVideo, AVI and other common video le types Web le types: HTML, PHP, ASP, DHTML, XML, Java Script, CCS, JSP, CFM, SGML and other common web le types Multimedia formats: Shockwave, QuickTime, Director, Authorware, Flash, VB application, IPix, etc Databases: Oracle, MS Access, FileMaker Pro Desktop application source formats: MS Word, MS Excel, MS PowerPoint, MS Project, MS Visio, Adobe Pagemaker, Adobe Illustrator Publishing formats: PostScript, PDF, TIFF Other common le types: TXT, EXE, ZIP, RTF, CSV, DOT, etc.
Figure 8.1 Standards and Formats in a Digital Objects Management System

(From McKnight and Livingston, 2003, p.9)
In one weeks crawl of Danish web sites by the Danish Royal Library in 2003, the following formats were identied from a total of 688 029 documents: document-like formats: 3 image formats: 6 sound formats: 3 video formats: 3 conguration and metadata formats: 2 program formats: 2 unknown format: 1.
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Of the total, 66.78 per cent were text les in HTML format, 19.17 per cent were image les in GIF format, and 10.12 per cent were image les in JPEG format; that is, 96 per cent were in HTML, GIF, or JPEG formats (Clausen, 2004, p.5). The problem posed by the multiplicity of formats (one estimate is more than 15 000 (Clausen, 2004, p.4)) is compounded by their nature. Many formats are proprietary, that is, they are the property of an owner who, for commercial reasons, is not willing to provide access to documentation about them, and who may require a fee to be paid for their use. Because one of the essential requirements of nearly all digital preservation strategies and techniques is a thorough understanding of le formats, the lack of access to full documentation about proprietary formats presents a major barrier. By comparison, the documentation for open formats, those that are in the public domain, is much more accessible. Consequently open formats are considered much more favourably for use in digital preservation applications than proprietary formats. To illustrate this point, consider that Microsoft Ofce documents (such as Word and Excel les) can be preserved as PostScript, PDF, DSSL, RTF, ASCII, SGML, TIF, CGM, or PNG les, each of which have different le format specications. Of these, PostScript, PDF, and RTF are proprietary le formats (Moehle, 2004). A report on the preservation of E-Prints (digital duplicates of academic research papers that are made available to improve access to these papers) commissioned by JISC in the UK devotes six of its 16 recommendations to le formats: 1. 2. 3. 4. 5. 6. Recognise the preservation risks of le formats Adopt open, standards-based le formats Investigate the use of XML formats to describe data and metadata Plan for migrating rare and obsolete le formats Maintain le format information Include le format identication functionality in E-Print repository software (James et al., 2003, pp.89).
Four main responses to the issues posed by the proliferation and complexity of le formats can be identied: le format registries, standardizing le formats, restricting the range of le formats handled in digital preservation systems, and developing archival le formats such as PDF/A.
File format registries

One response to the issues presented by le formats that is gathering momentum is the establishment of le format registries (also called repositories or depositories). Although these have existed for some time in the computer science arena, it is only recently that registries have been established for digital preservation purposes. These registries exist to provide detailed and reliable information about le formats, the key to many digital preservation activities. In their inuential case study of le formats, Lawrence et al. note that the risks associated with migration, such as the differences between the source and target formats, are quantiable, so the risks of migration can be managed (Lawrence et al., 2000, pp.1213). The kinds of le format problems that can arise as a result of migration are illustrated in the report of the Representation and Rendering Project at the University of Leeds. For example, format migration difculties, where
migrating an object created in version 1 of Macromedia Freehand to versions 5 and 9 resulted in different letter spacing, word spacing and kerning, although the drawing part of the object remained intact; objects migrated from MacDraw to Illustrator were unreadable (University of Leeds, 2003, pp.69). Risks of this kind are minimized if le formats specications are available. The experiments reported by Lawrence et al. found that the most difcult aspect of this project was the acquisition of complete and reliable le format specications (Lawrence et al., 2000, p.13). Their study concluded that information about le formats that is publicly available is vulnerable because it relies on the efforts of interested individuals. They concluded that there is a pressing need to establish reliable, sustained repositories of le format specications, documentation, and related software, to support migration (Lawrence et al., 2000, p.13). The Representation and Rendering Project reached the same conclusions: accessible le formats information falls far short of what is required to successfully tackle the problems of data obsolescence and the accuracy of most of what was available is mediocre at best (University of Leeds, 2003, p.42). It specically recommended that, because le format information available on the web is vulnerable, it should, as a matter of urgency, be captured and preserved in a le format registry. There are now a number of le format registries for digital preservation, and research to develop tools to apply them to the automation of digital preservation is under way. Christensen, for example, describes how a format registry can be used as part of the automated processes of running a web archive, by such things as matching up the best viewer with a specic le format so that it is rendered as accurately as possible. He provides the example of HTML les rendered by Notepad, which are intelligible but far from perfect (Christensen, 2004). PRONOM is one le format registry set up specically to support digital preservation. It has been developed by the National Archives (UK) to provide and manage information about le formats and about the software applications used to render these formats. It was originally developed as an in-house tool to support the National Archives digital preservation activities, but is now accessible on the web. PRONOM makes publicly available specications of software and the le formats that each piece of software can read and write, and records the degree to which the content remains unchanged by different products, and the support lifetime of products used to create, view or update electronic records. This is a response to the short life cycle of software and to the fact that older le formats are not necessarily supported by later versions of software, or, if they are, only with alteration of formatting or content, a situation inimical to the faithful reproduction of electronic records. PRONOM also provides information about software vendors and product support providers. It plans in the future to provide more detailed information about individual le formats. (This section is based on the PRONOM web site (www.nationalarchives. gov.uk/pronom).) Another le format registry developed to support digital preservation is the Global Digital Format Registry (hul.harvard.edu/gdfr/), set up to be inclusive in coverage, detailed in representation, rigorous in validity, public in discovery and delivery, and sustainable over archival time-spans. It aims to minimize the duplication of effort that occurs with the maintenance of multiple local format registries. Established in 2002, it has developed use cases, which are available on the Registrys web site, to identify some potential situations where a format
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
repository would be used. These include assessing the risk associated with a digital format, collection audit, validating the ingest of a digital object with a new format, monitoring for technology obsolescence, identifying rendering conditions for a digital object, determining the appropriate migration path for a digital object, determining the format of a given digital object, and identifying the migration pathway for a format (Abrams and Seaman, 2003). Other le format registries used to support digital preservation include Wotsits Format: The Programmers Resource (www.wotsit.org) and My File Formats (www.myleformats.com). These have mostly been established and maintained by computer scientists or enthusiasts rather than by those involved in digital preservation. The le format registry response to le format issues attempts to deal with all le formats, making no distinction among them. Other responses are based on restricting the range of le formats, both those in which les are created and those which digital archives agree to manage and preserve.
Standardizing le formats
If good le formats are used for creating digital materials, the difculties of preserving these digital materials will be minimized. This is the thinking that lies behind the digital preservation approach centred on encouraging document creation in those le formats that are most likely to be sustainable over the long term. Investigations into the effectiveness of different formats for this purpose are ongoing. For example, in its search for ways to ensure sustained accessibility to authentic archival records in the long term the Digital Preservation Testbed at the Nationaal Archief of the Netherlands investigated the sustainability of different record types (text documents, e-mails, spreadsheets, databases) in different le formats MS Word or WordPerfect for text documents, Outlook, Eudora, Novell GroupWise, Hotmail, and Kmail for e-mails, MS Excel and Lotus for spreadsheets, and MS Access and Oracle for databases. Together these account for more than 90 per cent of Dutch Government records (Slats, 2004). What are the characteristics of good le formats? They are the formats that are most likely to be viable for longer periods and are most likely to be open source and widely available and supported. Clausen tells us that most le formats fall out of fashion within a few decades (Clausen, 2004, p.23) and suggests a list of criteria by which we can predict the ongoing viability of a le format. These include: Openness criteria: for example, open, publicly available specication for the format; specication is in the public domain; not encrypted Portability criteria: for example, independent of hardware; independent of operating system; independent of other software; independent of particular institutions, groups, or events; widespread current use; containing little builtin functionality Quality criteria: for example, low space cost; highly encompassing; robustness; simplicity; highly tested; loss-free (Clausen, 2004, pp.1112). Using standard le formats is not a long-term digital preservation strategy. It does not address the obsolescence of le formats, but merely slows down the rate at which it occurs. It is based on the assumption that, for popular le
formats, obsolescence and compatibility problems will be addressed because of the large number of users, who will require that software to handle and render this format be upgraded to be usable in new operating systems and on new hardware platforms. However, this strategy has a limited lifespan: it is suggested that reliance on standards may lessen the immediate threat to a digital document from obsolescence, but it is no more a permanent preservation solution than the use of gold CDs or stone tablets (Kenney et al., 2003). Identifying and applying standard le formats is common practice. One example is the use of TIFF (Tagged Image File Format) to create preservation master images in many digitization programs. TIFF is used in this context because it is open and stable, and has a large user base (UNESCO, 2003, p.127). Creating digital materials in standard formats can be encouraged in situations where some creators of the material can be inuenced, for example, in-house digitizing programs. Format standardization is suitable for use in situations where the le formats can accommodate all or most of the characteristics of the digital materials. Consider a document created in a current version of a standard ofce word-processor, which, when converted to ASCII, will lose formatting. While this may not be considered an important loss for some kinds of materials, it may be signicant for others. In this case another le format that retains formatting, such as PDF, could be considered. A point in favour of the strategy of selecting and applying standard le formats is that it is likely to slow the rate at which le formats become obsolete. If widely used and supported le formats, or basic formats such as ASCII, are selected, then the software available to encode, decode and render these formats is likely to be available for longer periods of time than those for less popular formats. This strategy is not suitable for materials where there would be a loss of characteristics that are essential for its understanding if a particular le format is selected (UNESCO, 2003, p.128). Not all commentators are convinced that this strategy has merit. Rothenberg argues that to force digital documents into current standard forms (even if this sacrices some of their functionality) and then translate them, when current standards become obsolete, into whatever standards supplant the obsolete ones . . . is analogous to translating Homer into modern English by way of every intervening language that has existed during the past 2,500 years. The fact that scholars do not do this (but instead nd the earliest original they can, which they then translate directly into the current vernacular) is indicative of the fact that something is always lost in translation (Rothenberg, 1999a, p.11). He concedes, however, that converting digital documents into standard forms, and migrating to new standards if necessary, may be a useful interim approach while a true long-term solution is being developed (Rothenberg, 1999a, p.12).
Restricting the range of le formats

A strategy related to restricting the range of le formats in which les are created is for the digital archive to restrict the range of le formats that it agrees to receive and manage. The archive may convert material it receives into these
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
formats. This process is a formalized implementation of reliance on standards in which all digital objects of a particular type (e.g., color images, structured text) are converted into a single chosen le format that is thought to embody the best overall compromise amongst characteristics such as functionality, longevity, and preservability (Kenney et al., 2003). This practice has a long history in data archives. One example frequently cited is the UK Archaeology Data Service (ads.ahds.ac.uk), which prescribes the range of formats in which it will accept material. This strategy is most suitable where the digital archive is responsible for digital materials that are uniform in structure and content. The archive may be able to inuence the creation of these materials. For example, a government archives may specify and encourage, or enforce, that records created by its agencies are in the only standard formats that it will accept. Requirements for the successful operation of this strategy include clear submission guidelines, effective data conversion processes (if data conversion is needed), and quality control checking to a high standard (UNESCO, 2003, p.132133). Clausen proposes this strategy for web archiving. He recommends that the le formats harvested and added to the web archive should be constantly monitored so that action can be taken when new formats become widespread and old formats fall out of favour. He also suggests that the les received should be retained in their original format as well as in the converted and migrated versions to allow for higher-quality conversion or emulation at a later stage (Clausen, 2004, pp.2324). The use of stable, open le formats has the same advantages as noted in the previous section. The le formats selected will adhere to the characteristics, noted above, of openness, portability, and quality. In particular, the strategy has the potential to use the resources available for digital preservation in the most efcient way. This will ensure that resource requirements, such as specic software and expertise are minimized, and that the need for customized attention during migration processes is lessened. It is most suitable for materials for which content retention is essential but formatting and other characteristics are less signicant. While this strategy reduces the range of problems needing to be managed (UNESCO, 2003, p.132), it has the major disadvantage of not accommodating new varieties of digital expression (Hedstrom, 1998, p.195).
Developing archival le formats

The fourth response to the issues posed by the proliferation and complexity of le formats is the development of le formats with characteristics that will ensure their sustainability. One of these is PDF/A. Although PDF (Portable Document Format) is a proprietary le format, it has been developed as a format suitable for archival use. PDF, which reproduces the visual appearance of the digital material, is in widespread use. Because it allows the physical appearance as well as the content of digital materials to be reproduced, it is considered to be especially well suited for situations where authenticity is required. However, PDF is not an archival format. Although it is a proprietary format, Adobe (its owners) have published its specications, but are under no obligation to continue to do so. Some features of PDF, such as allowing parts of the document (such as fonts) to be dened outside the format, are incompatible with long-term sustainability or authenticity. In order to ensure the sustainability of PDF over long periods of time, an International
Joint Working Group was established under the auspices of the International Organization for Standardization to prepare an archival version of the PDF format. A draft standard for PDF/A, ISO 190051 Document Management: Electronic Document File Format for Long-Term Preservation Part I: Use of PDF 1.4 (PDF/A-1), was circulated for comment late in 2003. The PDF/A standard aims to dene a format that will ensure that PDF documents can be rendered consistently and predictably in the future. As an ISO Standard, PDF/A is more likely to be supported by publicly funded archival institutions than the proprietary PDF format. It is anticipated that the PDF/A standard will be nalized in 2005 (Sullivan, 2004). Differences between PDF and the draft PDF/A specications include audio and video content being forbidden in PDF/A, Javascript and executable le launches are forbidden, requiring all fonts to be embedded in PDF/A, the denition of colours in a device-independent way, and the forbidding of encryption (LeFurgy, 2003). The idea of a platform-independent format that would address the issues posed by the wide range and rapid obsolescence of le formats for digital media, especially video formats, was promoted in the mid to late 1990s by US radio station WGBH. A Universal Preservation Format especially suitable for media le types (particularly video) was advocated. This wraps the content in selfdescribing metadata that includes the technical specications required to build software to view the content. However, there has been little recent activity. More information is available on the Universal Preservation Format web site (info.wgbh.org/upf/index.html).
XML
XML (eXtensible Markup Language) is being actively investigated as an archival data format. XML is non-proprietary, being based on ISO88791986 (a standard for SGML, the Standard Generalized Markup Language), and is considered to promise long-term sustainability. XML is not, strictly speaking, a le format; rather, it is a set of rules to describe data and documents, a mark-up language in the same sense that HTML, in common use for web-based documents, is. XML was specically designed to be used regardless of the hardware platform and is supported by many open software applications. One principal reason for XMLs favourable consideration as an archival format is its stability and longevity. It has been used since the 1980s and there is a considerable amount of knowledge about it in the IT industry. The National Archives of Australia has adopted XML because even if the IT industry replaces XML with another data format in the future, we will still be able to create our own XML tools for as long as we wish because all the information needed to construct XML tools is publicly available (Heslop, Davis and Wilson, 2002, p.18). The National Archives of Australia is converting digital records in proprietary data formats into equivalent formats in XML. It has dened a number of XML schema for common formats (for instance JPEG and PNG) and these are available on its web site. It has developed open source tools that convert digital materials in some proprietary formats into XML. The process of using XML for digital preservation usually starts with analysing and tagging data so that the functions, relationships and structure of specic elements can be described, a process known as normalization (UNESCO, 2003, p.129). The UNESCO Guidelines provide three examples: a large-scale
Migration
147
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
project to apply XML tags to e-mails at the San Diego Supercomputer Center, XML tagging of e-mails and other common record formats by the National Archives of Australia, and XML representation of database tables by the VERS project at the Public Record Ofce Victoria of Australia (UNESCO, 2003, p.129). Another example is the use of XML as the basis of the DiVA (Digitala Vetenskapliga Arkivet) publishing system developed and used by a number of Scandinavian universities to establish and maintain electronic documents. In DiVA all metadata and contents of documents are converted to an XML-based format, which is an essential part of this systems long-term preservation strategy. Templates to convert documents created in word processors such as Word and Open Ofce to this format are provided (Mller, 2004). XML was recommended as the basis of an archive to store content originating as publisher SGML or XML les; the reasons included that there is a wide range of tools available for XML (Inera Inc., 2001). XML was investigated in the Dutch Digital Preservation Testbed study From Digital Volatility to Digital Permanence: Preserving Email (Digital Preservation Testbed, 2003) where, in a comparison with migration and emulation, it was found to be the most suitable preservation strategy for e-mails. One of the issues currently perceived with the XML approach is that it is based on close denition of elements of the data structures: the essence (see Chapter 5) has to be dened in advance. While this works well for simple data structures such as e-mails, it is not easy to see how all of the characteristics and behaviours of more complex kinds of materials can be captured and sustained. It is, therefore, likely to be applied primarily to structured or semi-structured data or documents for which retention of content, semantics and relationships is more important than any particular display characteristics (UNESCO, 2003, p.130). In other words, XML allows content to be reproduced, but is less accurate in representing the original formatting and layout of the document: it is not a one-size-ts-all solution (Aschenbrenner, 2004). However, as such documents constitute a very high proportion of the kinds of digital materials in some contexts, such as recordkeeping environments, this approach is likely to be fruitful. XML is at the heart of the persistent archives approach, as described by Thibodeau (2002). This approach is based on describing the essential characteristics of objects in a way that is independent of any specic hardware or software. Thibodeau describes it thus: In effect, tags delimit atomic preservation units in physical storage . . . Every tag is linked to one or more higher-level constructs, such as data models, data element denitions, document type denitions, and style sheets . . . research has shown that XML is currently the best method for tagging and articulating requirements at the information level and, to some extent, at the knowledge level (Thibodeau, 2002). The persistent archives approach is still experimental.
Migration
We rst need to establish what migration refers to in the context of digital preservation. Is it format migration, software migration, or version migration? In its simplest denition migration is:
to copy data, or convert data, from one technology to another, whether hardware or software, preserving the essential characteristics of the data (Kenney et al., 2003). A longer denition that is commonly encountered (for example in its use by Howell (2001, p.14) and Kenney et al., (2003)) is that migration is a set of organized tasks designed to achieve the periodic transfer of digital materials from one hardware/software conguration to another, or from one generation of computer technology to a subsequent generation. The rst use of this denition appears to have been by the Task Force on Archiving of Digital Information (1996). The term migration is sometimes used in the same sense as refreshing (dened earlier as the copying of data onto new media), but it is considerably more than that. Unlike refreshing, which is carried out because of concerns about obsolescence of the physical carrier, migration is additionally concerned with obsolescence of data formats and attempts to ensure that le formats remain readable. It is included in this chapter, rather than in Chapter 7, because it is based on ensuring that the digital materials remain accessible in the current technology, and are modied as and when this technology changes. Kenney et al. (2003) tell us more: The purpose of migration is to preserve the integrity of digital objects and to retain the ability for clients to retrieve, display, and otherwise use them in the face of constantly changing technology . . . Migration includes refreshing as a means of digital preservation but differs from it in the sense that it is not always possible to make an exact digital copy or replica of a data base or other information object as hardware and software change and still maintain the compatibility of the object with the new generation of technology. Migration theoretically goes beyond addressing viability by including the conversion of data to avoid obsolescence not only of the physical storage medium, but of the encoding and format of the data (Kenney et al., 2003, quoting in part from Task Force on Archiving of Digital Information, 1996, pp.ii,5). Migration has many varieties. Thibodeau, for example, notes simple version migration, format standardization, typed object model conversion, Rosetta Stones translation, and object interchange format (Thibodeau, 2002, p.23). The most commonly used migration process is simple version migration: migration within the same family of products or data types, for example versions of WordPerfect (.wpd les) or Excel (.xls les). He notes that software vendors provide this kind of migration and there is, therefore, the likelihood that support will be discontinued when such products are withdrawn (Thibodeau, 2002, p.23). Another migration process is format standardization, noted elsewhere in this chapter as a prerequisite or favourable pre-condition for successful migration over time (Thibodeau, 2002, p.23). Wheatley discusses in more detail the different kinds of migration and offers an extended example taken from his work on the Domesday Project (Wheatley, 2001). He notes that migration has to date been concerned mainly with relatively simple digital material and that maintaining access to this material has largely
Migration
149
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
occurred by applying format migration. What happens, he asks, with complex digital material? The denition in the Task Force on Archiving of Digital Information 1996 report (noted above) indicates a set of organised tasks and Wheatley denes these tasks in terms of the preservation techniques currently available. He identies these as: Minimum preservation preserving a copy of the bit-stream (noted in this chapter as bit-stream copying and refreshing) Minimum migration migration that uses simple manual or automated techniques to improve human viewing; for example, removing all characters except common ASCII characters from a word-processed le, resulting in a text le that is easy to access but no longer has any formatting or structural information Preservation migration, broken down into Basic preservation migration recording screen shots of the software in use; for example, a snapshot of the screen display on the obsolete platform Annotated preservation migration basic preservation migration, with the addition of textual descriptions describing the original materials function and look and feel Complex preservation migration annotated preservation migration with the addition of descriptive information about the original material; for example, video sequences captured to record key processes such as running the application, scrolling through content Recreation recoding the digital material by hand; for example, retyping the text in a current application and adding formatting to match the original Human conversion migration migration with elements of the original incorporated into the nal migrated material; for example, text and image data from the original is incorporated into the migrated object Automatic conversion migration uses software tools to interpret digital material and modify it into a new form; for example, a le from obsolete wordprocessing software is output as a current Word document. Wheatley notes this as a good example of a traditional view of migration. Wheatley illustrates each of these tasks with examples from the digital materials which were the test cases in the CAMiLEON Project. He examines the advantages and disadvantages of each and the possible situations where each is best applied. For instance, he suggests that the minimum migration approach has the advantages of simplicity and ease of execution, and although all functionality is lost it provides a cheap and reliable way of getting to a substantial amount of the intellectual content. It has potential for application as a useful stop gap measure as long as the original bit-stream and documentation are also maintained. He concludes: Although much recent work has gone a long way to establish the terminology that provides a common language for the discussion of preservation issues, the description of migration terms in particular has remained relatively confused. To continue progress in the development of practical preservation strategies we should not be afraid to introduce new and more specic language or to redene ambiguous terms of old (Wheatley, 2001).
No matter how it is dened, migration has been, with emulation, the principal preservation technique applied to date. It has a long history. Granger notes that it has been the only serious candidate thus far for preservation of large scale archives (Granger, 2000), and there is considerable expertise on the part of data administrators and computer scientists in migration. The literature of digital preservation includes some substantial studies of migration. One is the report Preserving the Whole: A Two-Track Approach to Rescuing Social Science Data and Metadata (Green, Dionne and Dennis, 1999), a detailed case study of preserving data in obsolete column binary format, and its associated documentation. This case study is based on the experience of the data archiving community, in particular that of Yale University Librarys management of social science numeric data since 1972. It identied that the best strategy for data in column binary format was to convert it to ASCII, because this format is software independent and also preserved the original content. This approach did, however, require a le-by-le migration strategy that is time-consuming for large numbers of les (Green, Dionne and Dennis, 1999, p.24). One conclusion of this study was that the existence of detailed documentation about the obsolete column binary format meant that there were many options available to migrate this format to others and also to read data in it (Green, Dionne and Dennis, 1999, p.[vi]). A Dutch study of migration investigated the issues of large-scale migration, in the order of 100 terabytes (Van Diessen and Van Rijnsoever, 2002). This quantity of data poses major resource issues: Lets say you want to migrate all this information to an optical storage medium. Current optical storage media have a capacity of around 5 GigaBytes and a write speed of around 4 MegaBytes/second. A quick calculation shows that a complete migration to optical storage would take at least 290 days (100 TeraBytes/4 MegaBytes per second)! (Van Diessen and Van Rijnsoever, 2002, p.v). This report developed medium migration indicators which, when applied in conjunction with changes in the system architecture or load characteristics, can be used to ascertain when migration needs to be carried out. The report also notes the actions required to manage media migration effectively in an electronic deposit system (Van Diessen and Van Rijnsoever, 2002). The National Library of Australia has long experience of migration. In 1998 it surveyed some of its material stored on disk and reported on the practical aspects of migrating oppy disks to CD-R (Woodyard, 1998). The aim was to migrate Australian publications that existed on oppy disks to CD-R. The data on a sample of disks, chosen to represent variables such as operating system, software, disk size, and age, was transferred to a network drive, downloaded to the hard drive of a PC containing disk writing software, and then written to CD-R. This study identied problems such as lack of the appropriate hardware or software, disk errors picked up by disk scan software, and illegal characters in le names. This trial highlighted the need to take further actions, including copying existing oppy disk publications selected for preservation to CD-R as soon as possible, establishing ongoing copying procedures for oppy disks, developing guidelines for selecting material to be preserved, and continuing trials of other forms of migration. The National Library of Australia has also
Migration
151
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
investigated digital materials in its manuscript collections. Here some issues assume greater signicance for example, the material is often old by the time it reaches the Library, and there may be a need to maintain authenticity of features such as layout (crucial to the understanding of poetry and some other literary works) (Woodyard, 2001). The Library established a Migration Strategy Group to develop further understanding of the issues and to draw up strategies for migration. One of its activities was a migration of PANDORA les to identify dead and depreciated HTML tags in the archive, which may not be recognised by later browsers (Webb, 2000c) and to see what the issues were in migrating to a later version of HTML. The outcome has suggested that there are likely to be scalability issues if this is carried out on a large scale (for example, 127 dead HTML tags were found). Software used by data recovery companies to determine older le formats and write translation tables has been purchased and used in the Librarys migration exercises. The National Library has recognized that these methods of migration are not a guarantee of preservation, but considers that they are necessary steps to increasing the period during which digital information remains accessible. Migration, as suggested above, is the preferred preservation strategy for most digital archives. It is favoured for the reasons already noted: the long history of its use, and our experience with it, at least for some kinds of materials and formats. There is some optimism that it will become simpler to apply as automated techniques are developed (Jones and Beagrie, 2001, p.102). However, the list of arguments against migration is long. Rothenberg, whose view is that migration is essentially an approach based on wishful thinking (Rothenberg, 1999a, p.15), trenchantly summarizes these when he characterizes migration as labor-intensive, time-consuming, expensive, error-prone, and fraught with the danger of losing or corrupting information (Rothenberg, 1999a, p.13). Migration needs to occur on a regular basis throughout the life of the materials, each time different in that it requires analysis of data structures, rules to be developed to control the transformation from one data format to another, programs to be written, and quality control procedures (UNESCO, 2003, p.139). Migration, therefore, has signicant ongoing costs associated with it. It is unlikely to benet from increases in computing power, Rothenberg suggests, because it is highly dependent on the particular characteristics of individual document formats. He also considers it unlikely that automated migration techniques will assist; in fact, he encourages us to consider them with great suspicion because of their potential for silently corrupting entire corpora of digital documents by performing inadvertently destructive conversions on them (Rothenberg, 1999a, p.15). Complex materials are not, at present, considered to be migratable because the loss of functionality and compromising of integrity and other attributes of the materials will be outside acceptable ranges (Beagrie, 2001, p.104). Finally, Hedstrom notes that despite our long experience with migration, we have no reliable data about the costs it will incur (Hedstrom, 1998, p.198). Version migration is worth noting in more detail. The backwards compatibility of software can be used to migrate digital materials: most applications can read older le formats created in earlier versions of that software. Relying on this for migration means depending on software vendors to develop and maintain their products ability to read earlier formats, and experience shows that this is not often the case (University of Leeds, 2003, p.4). Jones and Beagrie note that, although backwards compatibility is offered by an increasing number of vendors,
its continued availability is dependent on market forces which are notoriously volatile [and] it may therefore cease to be available with little or no warning (Jones and Beagrie, 2001, pp.107108). The UNESCO Guidelines distinguish between backwards compatibility and version migration: for the former, the presentation may be limited to temporary viewing; by comparison, version migration permanently converts documents into a format that can be presented by the current version of the software (UNESCO, 2003, p.137). Common examples are word processing software and web browsers, which can usually interpret and display material created in earlier versions of that software or in earlier versions of HTML. The National Archives (UK) has decided that In the case of passive systems, we are quite convinced that migration is the only way to go. However, we intend to use migration sparingly; our view is that we should only migrate a system when it becomes absolutely necessary; that is to say when it cannot be run on future generations of software or hardware. To take the example of Microsoft Word, we would only migrate documents written in Word version 1 if it ever became impossible to read them on the latest version of Word (Thomas, 2004). For migration to be successful certain requirements should be met. These include (based on Jones and Beagrie, 2001, p.104 and UNESCO, 2003, p.139): Written policies and guidelines Rigorous documentation of the procedures that were followed Strict quality control procedures, applied during the development stages and also after migration Retention of copies of the original material if some essential elements may be lost in migration Testing the migration process before it is fully implemented. Migration is best applied to less complex digital materials where the essential elements to be preserved are reasonably straightforward and do not depend on the look and feel of the material, and do not involve executable les and to collections of materials that are in uniform formats, such as collections of audio or digital image materials, especially if these formats are not proprietary (UNESCO, 2003, pp.139140). Some thought has been given to the functions that migration software should be required to perform. It should Read the source le and analyze the differences between it and the target format Identify and report the degree of risk if a mismatch occurs Accurately convert the source le(s) to target specications Work on single les and large collections Provide a record of its conversions (Lawrence et al., 2000, p.13). The aim of migration is to ensure that when it occurs, usually in response to specic triggers such as software upgrade or installation of new software, there is little or no loss in content or context (Jones and Beagrie, 2001, p.103). While some commentators envisage an increasing role for migration, especially for
Encapsulation
153
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
migration of simple data objects and for migration on demand (Wheatley, 2001), others suggest a diminished role as we move towards open standards and formats, and towards software with built-in format conversion functions (Dollar, quoted in Wheatley, 2001). Experimentation with migration continues, such as the on-access format migration experiments by LOCKSS in relation to web content (Rosenthal et al., 2005).
Viewers and migration on request

Migration on request that is, migration that is triggered by a request to access the digital material has been promoted principally by the CAMiLEON Project. The CAMiLEON Project (noted in more detail in Chapter 7) investigated migration and in the process came up with the concept of migration on request. Migration on request differs from other migration techniques because of the point at which the migration is applied. Migration is usually carried out when certain conditions apply, such as impending obsolescence of a data format or a new software installation, whereas migration on request occurs when a user wishes to access digital material. This material is maintained in its original bitstream, along with a migration tool (a conversion utility, or viewer) which is able to run on the current computing platform. Viewers are in common use; Microsoft, for example, provides free viewers that allow people to view, but not edit, documents produced in its Ofce software. This migration tool is maintained, perhaps through emulation techniques, so that it can always run on the current platform (Wheatley, 2001). This approach, which Clausen calls conversion on demand (Clausen, 2004, p.16), has two main advantages (albeit conjectural because migration on demand is still at proof-of-concept stage) over standard migration approaches. First, it reduces the possibility of cumulative errors (loss of functions or of content) typically introduced during each migration, so that loss of essential properties of the digital material and of the attributes which contribute to its authenticity are minimized. Second, it has the potential to reduce the costs incurred in regular migration. Because the migration is performed on demand when users request material, mass migration, with its signicant resource demands, is avoided. Against these advantages must be set some potential risks. One is that it may not be possible to develop viewers or software tools that work for complex materials. Another is that not all elements of the material will be re-presented satisfactorily by the viewers. And, of course, viewers and other software tools needed for migration on demand will themselves need to be migrated (UNESCO, 2003, p.143). Migration on demand needs further investigation before it can be added to the digital preservation toolbox. If it moves from its experimental status, it is likely to be of most use for digital materials for which there is low demand and for which the cost of regular migration is high.
Encapsulation
The concept of encapsulation has been on the digital preservation agenda for over a decade. Encapsulation is a technique of grouping together a digital resource and whatever is necessary to maintain access to it . . . [including]
metadata, software viewers, and discrete les forming the digital resource (Jones and Beagrie, 2001, p.108). In some cases, rather than including the actual software that reads the data in the bundle, metadata that points to the software at another location, or to the softwares specications is included. Encapsulation is widely used: for example data, metadata and software is often wrapped in XML metadata, thus providing the data in a way that can be understood by a wide range of technologies (UNESCO, 2003, p.131). This is the approach that has been taken in the VERS Project. The OAIS reference models AIPs (Archival Information Packages), described in Chapter 5, are another example of encapsulation. Encapsulation is a prerequisite for emulation. For emulation, according to Rothenberg, the information that needs to be encapsulated is of three kinds: the document and its software environment le(s) representing the original bit stream of the document, plus le(s) representing the original executable bit stream of the application program, plus le(s) representing the bit streams of the operating system a specication for an emulator for the documents original computing platform or a copy of or pointer to an emulator specication developed elsewhere explanatory material, labeling information, annotations, metadata about the document and its history, and documentation for the software and (emulated) hardware included in the encapsulation (Rothenberg, 1999a, pp.1718). The advantages that Jones and Beagrie perceive encapsulation has to offer as a digital preservation strategy are based on the way it groups together relevant information, as all supporting information required for access is maintained as one entity (Jones and Beagrie, 2001, p.108). Some skepticism is, however, evident in the statement that ostensibly, the grouping process lessens the likelihood that any critical component necessary to decode and render a digital object will be lost (Kenney et al., 2003). In theory, everything that is necessary to access the digital object, along with the digital object itself, is maintained into the future, but this is a risky approach to take because the only test of an encapsulated specication is at the point it is used to implement a rendering tool. The risk of missing vital information in the specication seems to invalidate this approach (University of Leeds, 2003, p.39). Encapsulation by itself is not considered a viable digital preservation technique because it does not really address the basic problem of technological change (UNESCO, 2003, p.131). The software encapsulated will still become obsolete. It is a key part of emulation, and is best considered as a strategy that is either a part of or a prerequisite for other digital preservation approaches. The UNESCO Guidelines suggest that encapsulation is best considered as a basic good practice for all objects, one that may facilitate other strategies (UNESCO, 2003, p.132).
Digital mass storage systems

Digital preservation activities require reliable techniques for storing large quantities of data. Storing digital objects leads, if nothing else, to large byte-counts,
Digital mass storage systems 155
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
note Linden et al., who point out that individuals now routinely deal with gigabytes of data for their personal archiving requirements (Linden et al., 2005, p.3). Although digital mass storage systems can be considered as basic to computer system management and, therefore, outside the scope of this book, they are included here because one of the fundamental routines of digital preservation is the archiving of large amounts of data. The digital preservation community has needed to learn about computer systems management as it develops. Concepts and practices such as ensuring security of computer systems, redundancy, RAID, total cost of ownership, are all part of the knowledge required to develop and maintain digital mass storage systems that will provide secure, long-term storage of both the digital materials and associated metadata. A digital mass storage system can be simple or very complex, depending on what is required of it. At its most basic, it is little more than a straightforward storage and back up system for a single type of le format, its derivative copies and related metadata, available for purchase off the shelf (International Association of Sound and Audiovisual Archives Technical Committee, 2004, p.49). The feasibility of a relatively inexpensive personal digital mass storage system for small institutions has been noted: these are technically feasible, but their practical implementation . . . is obstructed by the high initial investment and annual running costs mainly in the provision of software (Schller, 2000). At its most complex, a digital mass storage system is a comprehensive, fully automated system designed to store, manage, maintain, distribute and preserve an array of complex digital heritage objects and their related metadata (International Association of Sound and Audiovisual Archives Technical Committee, 2004, p.49). Typically it requires a high nancial outlay on hardware, software, and ongoing nancial investment in running, upgrading, and maintenance. Although the costs of hardware and storage continue to decrease, the costs of the software needed to run a digital mass storage system, such as database management software and integrity checking software, do not decrease. Technical specications for digital mass storage systems are available. The International Association of Sound and Audiovisual Archives Guidelines on the Production and Preservation of Digital Audio Objects (2004, pp.4950) provide brief descriptions of the specications of a number of large-scale systems, such as the National Library of Australias Digital Object Storage System, and systems at the National Library of Norway, Sdwestrundfunk (a German public broadcasting corporation), and ScreenSound Australia (now the National Film and Sound Archive), along with more detailed technical discussion. The Digital Preservation Coalition has published a Technology Watch Report on mass archival storage systems, based on experience at the British Library, which includes specications and notes lessons learned in the process (Linden et al., 2005). Calls for tender and similar documents available on the web are a further source of technical specications. As well as the technical requirements, digital mass storage systems require well-dened management processes, including allocating unambiguous responsibility for managing data storage and protection appropriate technical infrastructure to do the job system capabilities such as: sufcient storage capacity
demonstrated reliability ability to manage redundant storage (as digital media has a small, but signicant failure rate, redundant copies of les at every stage are a necessity) error checking means of storing metadata and of reliably linking metadata (International Association of Sound and Audiovisual Archives Technical Committee, 2004, pp.5152). Commercial data archiving services exist. One such is dataVault, whose services provide small businesses with a reliable data protection service. Its secure server is located in a state of the art data center facility which features extensive security features such as re suppression system, backup power supply, environmental control, surveillance, and full systems redundancy [Redundant IBM RAID 5 servers], including redundant network connections from Tier One providers (Global Data Vault Inc., 2005). dataVault provides a web-based facility for backing up data and monitors these backups and advises clients if errors occur But can commercial services such as dataVault be considered as trusted digital repositories, in the sense that these are described in Chapter 5? The dataVault web site describes how the les it stores are compressed and restored on the y. But digital preservationists would do well to be wary of such practices and raise questions about them. What are the implications of this for authenticity, for instance?
Conclusion: combining principles, strategies and practices

How do all of the various principles, strategies and practices noted in this chapter and the two previous chapters t together? The key lies in their effective combination. There is, as yet, no single recommended combination, but we can identify some that seem likely to rise to the top of what on the surface seems a muddy pond. Collections of digital materials that consist of only a single simple data format, or a very limited range of formats, can conceivably be made sustainable by applying one strategy or technique. Diverse collections require a range of solutions. As well as addressing better the range of attributes that need to be preserved, there are potential benets of redundancy in pursuing more than one strategy: even with good planning, a single strategy may fail leaving the programme with no means of access (UNESCO, 2003, p.149). Combinations already in use are: Standard formats and migration: for example, TIFF is often chosen as a standard le format for collections of images to prepare for migration in the future Encapsulation and emulation go together, as noted in the section above Standard data formats (PDF), viewers and migration (of data and metadata encoded in XML) are combined in the VERS strategy.
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Chapter 9
Digital Preservation Initiatives and Collaborations
Introduction
Preservation is a problem domain that demands collaborative action (Ross, 2004) How can we begin to describe the wide variety and large number of digital preservation initiatives? What is a useful starting point to reect on digital preservation activities in a way that will allow us to benet from the experience of these initiatives? One approach is to locate and apply existing typologies, but in the rapidly changing landscape of digital preservation these are likely to be outdated by the time they become public. This chapter rst considers the theme of collaboration that is strikingly apparent in many digital preservation initiatives. It then notes some ways in which these initiatives have been categorized and describes a number of digital preservation activities in those categories, which are structured rst around geographical considerations international, regional, national or sectoral and then subdivided into services (projects actually attempting digital preservation) and alliances (collaborations to develop approaches to digital preservation). Only a selection of initiatives and collaborations is covered here. Their description is intended to illustrate the range and nature of current digital preservation activities, not to provide a comprehensive list.
Collaboration
One characteristic of the digital preservation agenda is that collaboration is rmly embedded in it and has been so from the earliest days of its compilation. Partnerships have always been important in the digital preservation community. From the very beginning it was apparent that no one
158 Digital Preservation Initiatives and Collaborations
organization whether library, government or academic could adequately archive, preserve and continue to provide access to the digital material, even with stringent selection criteria (Hodge and Frangakis, 2004, p.63). Certainly collaborative activities are more and more evident at all levels, and collaboration is seen increasingly as one of the keys to effective confrontation of what seem at times to be overwhelming threats posed to digital preservation. A report of a recent workshop notes that one key theme running throughout the day was the need for active collaboration at every level and across sectoral and geographic boundaries. Speaker after speaker illustrated how this collaboration was essential (Digital Preservation Coalition, 2004). The topic of Article 11 of the UNESCO Charter on the Preservation of the Digital Heritage (UNESCO, 2004) is Partnerships and cooperation and the UNESCO Guidelines devote a whole chapter to collaboration (UNESCO, 2003, chapter 11, Working Together). Increasing collaboration at local, national, regional and international levels can be readily observed, and the rest of this chapter describes some examples of collaboration. The reasons for the prominence of collaborative action in digital preservation are to be found in large part in the scale of the issues and uncertainty surrounding how to address them. Because digital preservation is expensive and resources are scarce, collaborative activities can enhance the productive capacity of a limited supply of digital preservation funds, by building shared resources, eliminating redundancies, and exploiting economies of scale (Lavoie and Dempsey, 2004). The issues are similar across different kinds of organizations (such as libraries and archives) and different sectors (for example, different business sectors) so it makes sense to capitalise on the potential benets of pooling expertise and experience; there may also be political pressure to collaborate (Jones and Beagrie, 2001, p.38). Another compelling reason to collaborate is the uncertainty about where the responsibility for preserving digital materials lies. There are many stakeholders (noted in Chapter 2), none of whom can realistically expect to develop workable, scalable solutions on their own. Authors are encouraged to collaborate in archiving their own works in university-based digital repositories, for example. Journal publishers are collaborating with libraries or library-based organizations such as JSTOR and LOCKSS to provide continuing access to their publications. Morris sees promise in the way that organisations from different parts of the information chain are beginning to work together to address some of the problems (Morris, 2002, p.130), but cautions that the only way we can hope to nd reasonable and scholarship-friendly solutions is to work together. We have to make sure that there is close communication between the plethora of initiatives in different parts of the world, so that none of us wastes time and money re-inventing the wheel . . . And we need to make sure that all the members of the information chain information creators, information users, and all the intermediaries in between are involved in the discussions and in creating the solutions (Morris, 2002, p.132). However, as the UNESCO Guidelines put it, Collaboration costs (UNESCO, 2003, p.64). Collaboration has the potential to sidetrack the main outcome by
Typologies of digital preservation initiatives 159
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
diverting attention from local objectives. Considerable effort has to be put into ensuring that unambiguous agreements able to be accepted by all parties are reached (Jones and Beagrie, 2001, p.39). Despite these and other concerns, it is acknowledged that the benets outweigh the disadvantages. The Guidelines identify the benets of collaboration as: Access to a wider range of expertise Shared development costs Access to tools and systems that might otherwise be unavailable Shared learning opportunities Increased coverage of preserved materials Better planning to reduce wasted effort Encouragement for other inuential stakeholders to take preservation seriously Shared inuence on agreements with producers Shared inuence on research and development of standards and practices Attraction of resources and other support for well-coordinated programmes at a regional, national or sectoral level (UNESCO, 2003, pp.6465). Two studies of collaborative activities have come down strongly in favour of their benets, both actual and potential. Case studies carried out by ERPANET indicate that, in the broadcasting sector, strong sectoral cooperation leads to increased condence and in turn, increased activity and development of common knowledge and understanding and suggest that this collaboration will result in the development of more robust and nancially viable tools and solutions (Ross, Greenan and McKinney, 2003, p.8). One major nding of Hodge and Frangakiss study of providing permanent access to scientic information in digital form was that developing partnerships is of increasing importance. The stakeholders in this arena have increasingly realized that partnerships are the only way to ensure that digital information will be preserved. As well as providing the possibility of preserving most, if not all, signicant scientic digital information, partnerships offer benets such as providing some measure of redundancy, sustainability and sharing of the cost for preservation (Hodge and Frangakis, 2004, p.3). An Australian example is the National Library of Australias strong commitment to international collaboration. In his 2002 survey of digital preservation activities, Beagrie observed that the National Library of Australia believes that international collaboration at many levels is essential in digital preservation (Beagrie, 2003, p.18) and noted the Librarys commitment to many international collaborations, including PADI with its international advisory group, a Memorandum of Understanding with the Digital Preservation Coalition, the Safekeeping Initiative with CLIR, the National Librarys role in the Conference of Directors of National Libraries, whose action plan includes aspects of digital preservation, and its participation in working groups such as the OCLC/RLG working groups on preservation metadata and on digital archive attributes (Beagrie, 2003, p.19).
Typologies of digital preservation initiatives

Collaboration takes many forms. At one level of categorization the distinction is made between internal collaboration (cooperation among sections in a large
organization) and external collaboration (cooperation with other organizations) (Jones and Beagrie, 2001, pp.3839). Beyond this, though, existing typologies are not particularly helpful in suggesting ways to group activities that acknowledge adequately the variety and extent of these increasingly fruitful collaborative activities. One typology takes an evolutionary approach by suggesting that collaboration comes at the end of a chain of activities: rst acknowledging that digital preservation is of concern at a local level, next acting by initiating a project, then consolidation by moving from projects to programmes, institutionalizing the activities to incorporate the larger environments, and nally externalizing the activities by embracing inter-institutional collaboration and dependency (Nancy Y. McGovern, in a presentation at the DPC Forum, London, 23 June 2004). Another typology is based on the nature of the organizations that carry out digital preservation: enterprise-based preservation services include research libraries, academic disciplines, and publishers; and community-based preservation services include JSTOR and The Internet Archive (Smith, 2003). The UNESCO Guidelines, which, as noted above, devote a whole chapter to cooperation, suggest four ways in which digital preservation collaborations can be categorized. In the centralized distributed model one partner takes the main role and other partners contribute in specically dened ways. While this model offers some cost sharing, develops expertise among partners, and may offer economies of scale, it does not necessarily encourage ownership by lesser partners. This model is probably good for beginning programmes seeking to collaborate with large, advanced programmes. More equally distributed models, with several partners who have similar responsibilities and commitments, may have issues of leadership and consultation may become time-consuming. This model is probably suitable where there are a number of players willing to share responsibility but none wanting to lead a programme. Very highly distributed collaborations have a number of participants who largely take responsibility for their own activities. This model may be a useful starting point for a preservation programme, raising awareness and allowing some steps to be taken and does not require high costs to participate. Standalone arrangements, curiously, are noted in the UNESCO Guidelines as the fourth model of collaborative activities. They allow the development of expertise, strategies and systems before collaborative partners are sought (UNESCO, 2003, p.67). The UNESCO Guidelines also provide useful information about how to make collaborations work effectively and how to set them up (UNESCO, 2003, pp.66,68). The typology applied in this chapter is loosely based around geographical considerations international, regional, national or sectoral and subdivided into services and alliances. These categories, based on geography and the nature of the collaboration, at times require arbitrary distinctions to be made. Although some of these programmes are categorized in this chapter as national, or local, and so on, such is the collaborative nature of digital preservation activities that their lessons are typically made available to a wider audience and are keenly scrutinized. Despite its imperfections this typology is applied because it accommodates most current digital preservation activities. Initiatives and programmes that have been described elsewhere in this book are only referred to briey in this chapter. Among other useful lists of digital preservation activities is the Directory of Digital Preservation Repositories and Services in the UK (Simpson, 2004b). This lists digital preservation service providers and repositories in the United Kingdom offering digital storage services who are considered to have a commitment to
International initiatives and collaborations 161
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Services International The Internet Archive International Internet Preservation Consortium JSTOR DSpace LOCKSS OceanStore NEDLIB
Alliances UNESCO RLG PADI OCLC CAMiLEON
Regional
National
Sectoral
Koninklijke Bibliotheeks e-Depot National Archives (UK) Digital Archive Digital Curation Centre AHDS UK Web Archiving Consortium PARADISEC CEDARS
ERPANET Digital Recordkeeping Initiative Digital Preservation Coalition NDIIPP Digital Preservation Testbed JISC
Figure 9.1 Initiatives and Collaborations Described in this Chapter
long-term digital preservation, this latter criterion excluding some commercial data storage services. The UNESCO Guidelines provide another listing focusing on collaborative activities (UNESCO, 2003, pp.6869). The online erpaDirectory provides evaluations of digital preservation projects (www.erpanet.org).
International initiatives and collaborations

International in this chapter is applied to two types of programme. The rst type are either those digital preservation initiatives or collaborations that were established from their inception to be available or applicable to participants from any country, or those whose concern is with digital materials from anywhere in the world, without any geographical limitation, such as country or state, or sectoral limitation, such as research libraries. The second type are those initiatives or collaborations that were initially established with a geographical or sectoral limitation, but have since become available to the international community. The international services considered here are The Internet Archive, the International Internet Preservation Consortium, JSTOR, DSpace, LOCKSS, and OceanStore, and the international collaborations noted are those of UNESCO, RLG, PADI, OCLC, and CAMiLEON.
International services
The Internet Archive (www.archive.org)
The Internet Archive, established in 1996 by Brewster Kahle, is a non-prot organization based in San Francisco whose aim is to provide permanent access
to digital materials, primarily those on the web. Its web site (on 4 January 2005) describes its aims as working to prevent the Internet a new medium with major historical signicance and other born-digital materials from disappearing into the past and promoting the ideals of open and free access to literature and other writings which have long been considered essential to education and to the maintenance of an open society. The Internet Archive is funded by donations from individuals and philanthropic organizations (among them the Hewlett Foundation, the Sloan Foundation, the Kahle/Austin Foundation, and the National Science Foundation) and by contract work it undertakes for bodies such as the Library of Congress, the national archives of the US and Britain, and the Bibliothque nationale de France. The Internet Archive contained over 400 terabytes at November 2004, the result of web crawls of all publicly accessible material on the web every two months plus, from 1999, targeted web crawls, some commissioned by specic organizations. This has resulted in several collections of web sites: The UK Central Government Web Archive: selected UK Government websites from 2003, collected for the National Archives (UK) Election 2002 Web Archive: almost 4000 web sites relating to the 2002 US elections, collected for the Library of Congress September 11th: archived web sites relating to the events of 11 September 2001 in the US, collected for the Library of Congress Election 2000: web sites relating to the US elections held in 2000, commissioned by the Library of Congress Web Pioneers: web sites illustrating the early years of the internet. The Internet Archive is not, as is popularly believed, a complete depiction of the web. It does not capture password-protected sites, dynamically-generated content, and other material. It justies its inclusive approach of archiving the entire publicly-available web by arguing that the cost of selection is greater and riskier than capturing all. The thorny question of intellectual property rights is addressed by collecting all publicly available material, but responding to requests for privacy and blocking access if a sites owner requests this. Technically, the Internet Archive is based on a large number of personal computers with IDE hard disks. The data is stored on DLT tapes and hard drives in ARC le format, a format being promoted as a standard for archiving internet material by Alexa Internet, who supplies the Archive with its web crawls. (Specications for the ARC le format are available at pages.alexa.com/ company/arcformat.html.) Copies of the collections are maintained at several sites, currently in Egypt and Holland. The data is migrated as appropriate. To counter data format obsolescence the Archive intends to collect appropriate software and emulators. The Internet Archive has until recently been primarily interested in capturing and storing web sites, but is now turning its attention to providing access. The Wayback Machine, launched in 2001, provides access to archived versions of web pages. The Internet Archive is active in research and development, for example to develop better web crawlers and better user middleware. It actively seeks collaboration, such as in the International Internet Preservation Consortium (netpreserve.org), formed in 2003 (described below), and in the Open Access Text Archive, a collaboration with libraries from many countries, which is
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
putting digitized books into open-access archives. In December 2004 over 27 000 books were available, with another 50 000 expected in the rst quarter of 2005; the goal is one million books. The Internet Archive has inuenced digital preservation by demonstrating that large quantities of web materials can be archived over time. (This section is based on Lyman and Kahle (1998), Smith (2003) and the Internet Archives web site (www.archive.org).)
The International Internet Preservation Consortium (netpreserve.org)

The International Internet Preservation Consortium (IIPC) was formed in 2003, initially for three years. Its members are the national libraries of Australia, Canada, Denmark, Finland, Iceland, Italy, Norway, and Sweden, the British Library, the Library of Congress, and the Internet Archive, with overall coordination by the Bibliothque nationale de France. The goals of the consortium are: To enable the collection of a rich body of Internet content from around the world to be preserved in a way that it can be archived, secured and accessed over time To foster the development and use of common tools, techniques and standards that enable the creation of international archives To encourage and support national libraries everywhere to address Internet archiving and preservation (netpreserve.org). It has six working groups: Framework, Researchers Requirements, Access Tools, Metrics and Testbed, Deep Web, and Content Management. To date it has provided two reports available on the Consortiums web site: one that presents a taxonomy of the challenges that web crawlers encounter when trying to copy content for web archiving, and another that classies the conditions encountered on web sites (such as static HTML documents, forms, JavaScript) and describes the issues that arise in harvesting these. The IIPC is developing software tools for web archiving and this has resulted in products such as Heretrix (an archive quality web crawler) and DeepArc (software that extracts database content to XML at les). The aim is to develop a toolkit of web archiving software that is open source and easy to install. If it achieves this aim, our ability to archive web content will be considerably enhanced. (This section is based on information from the International Internet Preservation Consortiums web site (netpreserve.org) and from comments made during the Archiving Web Resources International Conference, Canberra, 911 November 2004.)
JSTOR (www.jstor.org)
JSTOR, a non-prot organization based in the US, was established in 1994, initially to address the problems libraries faced in providing storage space for long runs of scholarly journals. It now has two aims: to develop a trusted archive of important scholarly journals, and to provide access to these journals as widely as possible. Today, users of JSTOR from many countries can search and retrieve high resolution images of journal issues and pages: 2 224 participants in November 2004, of whom 909 were located outside the US in 86 countries. JSTORs content covers scholarly journals with a focus on the humanities and
social sciences: 457 titles and 16.6 million pages in November 2004 (statistics at 23 November 2004 come from the JSTOR web site). Each journal title is digitized from its rst issue as a 600dpi TIFF image and a text le for searching purposes. JSTORs preservation strategies focus on stable standard formats, data backup, and redundancy (Heterick, 2002, pp.114116). JSTOR is extending its activities into born-digital materials as one partner in the Electronic-Archiving Initiative (or E-Archive). This seeks to preserve the publishers versions (the source les) of e-journals, because these often contain material that is not presented in the online versions, for example high resolution image les. To date planning and discussions with publishers and libraries have occurred, a working prototype archive has been developed, a productionlevel archive is now in development, and verication and normalization procedures are being developed. JSTORs signicance for digital preservation lies mainly in its business model, which provides benets to all of its stakeholders publishers, libraries, and users. It is subscription-based, and charges an initial one-off fee to all subscribers, which supports the costs incurred in digitizing journals and managing the resulting les. This business model, Smith suggests, promises to be sustainable over time (Smith, 2003, p.21). (This section is based on Heterick (2002), Smith (2003) a presentation by Eileen Gifford Fenton at the DPC Forum, London, 23 June 2004, and the JSTOR web site (www.jstor.org).)
DSpace (www.dspace.org)
DSpace began life as a sectoral initiative, as an institutional repository for the material produced by staff of MIT (Massachusetts Institute of Technology). At an early stage in its development DSpace was made available to other institutions and has been adopted around the world. DSpace is an institutional repository designed to enable the capture, distribution and preservation of the intellectual output of MIT but always with a view to its adoption by, and federation with, other institutions (Greenan, 2003, p.8). Developed jointly by MIT Libraries and Hewlett-Packard, it was trialled within MIT from February 2002 and launched in September 2002. The benets of using DSpace, according to its publicity, include its making content quickly and widely available, and its indexing capabilities; preservation benets include Longterm preservation for a variety of digital formats including text, audio, video, images, datasets, and more A long-term stable URL that can be used in a citation to link to items in DSpace (DSpace Communications Kit, 2003). A signicant characteristic of DSpace is that it is an open source system. The source code for DSpace was made available under an Open Source licence in November 2002. These licences are promoted by the OpenArchives Initiative, whose aims are to develop protocols and interoperability standards and tools to promote interoperability among multiple repositories (see www.openarchives.org). Since November 2002 it has been downloaded by several thousand institutions, more than 1500 in 2003 alone (Greenan, 2003, p.8). DSpace distinguishes between bit preservation and functional preservation. Bit preservation
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
ensures that a le remains exactly the same over time not a single bit is changed while the physical media evolve around it. Functional preservation goes further: the le does change over time so that the material continues to be immediately usable in the same way it was originally while the digital formats (and physical media) evolve over time (DSpace FAQ, 2003). DSpace recognizes that faculty will create digital materials in a wide variety of formats to suit their own aims, and that repositories must, therefore, handle these formats. It accepts all forms of digital materials and denes three levels of preservation for le formats supported, known, or unsupported. Bit-level preservation is carried out for all three levels. Supported le formats (those that are open and archival, such as TIFF, SGML, XML, AIFF, and PDF) are functionally preserved using format migration or emulation techniques. Known le formats (popular proprietary formats such as Microsoft Word and Powerpoint, Lotus 12-3, and WordPerfect) will rely on the likelihood that third party format migration tools will be developed for them. Unsupported le formats (those about which little is known, such as unique software programs) will not have any functional preservation applied. DSpace developers are working on procedures for uploading processes which will convert unsupported or known formats to supported ones, and is also developing its capability to carry out regular format migrations. DSpace assigns to material submitted a unique identier, stores provenance information, maintains an auditable history and record of changes to the archive [and] provides persistent storage (Sullivan et al., 2004). MIT also encourages collaboration among DSpace participants. An early collaboration was the DSpace@Cambridge Project (www.lib.cam.ac.uk/dspace) from 2003. As well as implementing a repository at Cambridge University, this collaboration is intended to act as a UK model. In 2003 the DSpace Federation was established, of which all institutions that have implemented DSpace are members. The intention of the DSpace Federation is to share technical innovation, content, and services and to promote interoperability among institutional repositories to support distributed services, virtual communities, virtual collections, and cataloging. This will happen through activities such as sharing in the development and maintenance of the DSpace source code, and promoting the DSpace service and interoperability of archival repositories. In Australia, DSpace is implemented at the ANU (Australian National University), Canberra (dspace.anu.edu.au). In 2004 ANU tested DSpace as a platform for its digital repository, and was trialling a Google DSpace search facility. DSpace is also being trialled at the University of Melbourne to ascertain its suitability for maintaining and accessing resources located in the University of Melbournes Percy Grainger Museum (Sullivan et al,. 2004). DSpaces inuence in digital preservation is growing because it provides a framework in which academic libraries and archives can develop strategies and practices in a collaborative international environment. (This section is based on Barton and Walker (2003), Greenan (2003), Smith (2003), Sullivan et al. (2004) and the DSpace web site (www.dspace.org).)
LOCKSS (lockss.stanford.edu)
The LOCKSS (Lots Of Copies Keep Stuff Safe) project is noted in detail in Chapter 8. It is based on the well-established preservation principle of
redundancy (keeping multiple copies as a safeguard against loss). LOCKSS is signicant in digital preservation terms because it established the feasibility of replication and peer-to-peer polling using standard personal computers.
OceanStore (oceanstore.cs.berkeley.edu)
The peer-to-peer concept demonstrated in LOCKSS is also the basis of OceanStore, whose web site describes it as a global persistent data store designed to scale to billions of users. It provides a consistent, highly-available, and durable storage utility atop an infrastructure comprised of untrusted servers. OceanStores features include data protection through redundancy and through its cryptographic techniques. Any computer can join the OceanStore infrastructure, and users subscribing only to a single OceanStore service provider can access other OceanStore servers. Data is cached in OceanStore promiscuously: any server may create a local replica of any data object. This offers benets such as faster access. OceanStore claims that it can offer durability which exceeds todays best by orders of magnitude because digital material is stored on hundreds or thousands of servers and so can be readily reconstructed. Only a global-scale disaster, they claim, could disable enough machines to destroy the archived object. Pond, a prototype of OceanStore, is being developed. (This section is based on information available on the OceanStore web site (oceanstore.cs.berkeley.edu).)
International alliances
The examples of digital preservation services noted above are all based on collaboration to a signicant extent. Some international collaborations are different in that they are intended primarily to educate, inform, and lobby, rather than to establish services that carry out preservation of digital materials.
UNESCO (www.unesco.org)
UNESCO (the United Nations Educational, Scientic and Cultural Organization) has assumed a prominent role in promoting digital preservation. Its general conference in 2001 adopted a resolution which drew attention to the need to safeguard endangered digital heritage. A discussion paper prepared by ECPA (the European Commission on Preservation and Access) (de Lusenet, 2002) and further consultation, including ve regional expert meetings in Canberra, Managua (Nicaragua), Addis Ababa (Ethiopia), Riga (Latvia), and Budapest (Hungary) in 2002 and 2003, resulted in the 2004 Charter on the Preservation of the Digital Heritage (UNESCO, 2004). This statement of principles is for advocacy and public policy purposes. Its 12 articles describe the value of digital resources and their vulnerability, note that it is vital that these resources remain accessible, and stress the need for action. They also note the importance of taking a life cycle approach, afrm the need to consult stakeholders and to dene criteria for selection, emphasize the importance of appropriate legal and policy frameworks, stress the need for cooperation, and note the responsibility of UNESCO (UNESCO, 2004, pp.1417).
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
In conjunction with the Charter on the Preservation of the Digital Heritage, UNESCO contracted the National Library of Australia to develop guidelines for digital preservation. These Guidelines for the Preservation of Digital Heritage (UNESCO, 2003) were made available via the UNESCO web site in 2003. They were the product of a wide international consultation process with all levels of interest groups, from governments to individual experts. The preface to the Guidelines states that their intention is to introduce general and technical guidelines for the preservation and continuing accessibility of the ever growing digital heritage of the world and that they are intended to complement the Charter on the Preservation of the Digital Heritage. They are best seen as a guide to the questions that programme managers need to nd answers to and do not claim to address every technical and practical question that will arise in managing digital preservation programmes, relying rather on establishing and stating principles that can be applied (UNESCO, 2003, pp.7,11). The UNESCO Guidelines have already proved valuable to the digital preservation community for their comprehensive approach and statement of principles, as the frequent reference to them throughout this book attests.
RLG (www.rlg.org)
RLG, founded as the Research Libraries Group in 1974, is a not-for-prot alliance of libraries, archives, museums, and historical societies with strong research collections. RLG was established to provide a mechanism for collaborative action on the problems facing research collections. Preservation has always been a major interest of RLG, demonstrated most recently by its signicant digital preservation activities, particularly in advocacy and raising awareness and in standards development. RLGs advocacy and awareness-raising activities include publication since 1997 of the electronic journal RLG DigiNews. Perhaps its most inuential activity in digital preservation has been its participation, with the Commission on Preservation and Access, in the Task Force on Archiving of Digital Information. The report of this Task Force in 1996 (Task Force on Archiving of Digital Information, 1996) laid the foundations for much subsequent work in digital preservation. RLG surveyed the digital preservation needs of its members in 1998 (Hedstrom and Montgomery, 1999). RLGs role in standards development is similarly signicant. It participated in 1998 in the development of the OAIS Reference Model. It has participated in working groups, including: RLG-OCLC Working Group on Digital Archive Attributes, whose nal report was Trusted Digital Repositories: Attributes and Responsibilities (RLG/OCLC Working Group on Digital Archive Attributes, 2002) OCLC-RLG Preservation Metadata Working Group, which produced A Metadata Framework to Support the Preservation of Digital Objects (OCLC/RLG Working Group on Preservation Metadata, 2002) Preservation Metadata: Implementation Strategies (PREMIS) Working Group, another joint RLG-OCLC working group, reported in 2004: Implementing Preservation Strategies for Digital Materials: Current Practice and Emerging Trends in the Cultural Heritage Community (OCLC/RLG PREMIS Working Group, 2004).
Other RLG collaborations include a fruitful working relationship and strategic partnership with JISC since 1999 (Dale, 2004, p.20). RLG staff have also been on the advisory groups of many collaborative digital preservation activities, such as CEDARS and CAMiLEON. RLG is a founding member of the DPC and has worked with the National Library of Australia on PADI. (This section is based on Dale (2004) and on the RLG web site (www.rlg.org).)
PADI (www.nla.gov.au/padi)
PADI (Preserving Access to Digital Information) describes itself as a subject gateway to digital preservation resources and is considered by many as the essential starting point for digital preservation matters. It was established in 1997 by the National Library of Australia to share information about digital preservation and support activities in Australia and worldwide. PADI demonstrates the collaborative aspects that characterize many digital preservation activities: it is developed in cooperation with CLIR (the Council on Library and Information Resources) and the UK-based Digital Preservation Coalition, and has an international advisory group. PADI is described in more detail as Case Study 6 in the Appendix of this book.
OCLC (www.oclc.org)
The activities of OCLC in digital preservation sit uneasily with the distinction between services and alliances that is made in this chapter. OCLC has been active in both arenas. Some of its collaborative activities with RLG are referred to above. OCLC was founded, as the Ohio College Library Center, in 1967, primarily to offer cataloguing services to US libraries. Now renamed Online Computer Library Center, it offers a very wide range of services, including digitization and preservation services, to an international clientele. In addition to the collaborations with RLG already noted, OCLCs digital preservation activities include a Registry of Digital Masters working group (again with RLG), and participation on the board of METS (Metadata Encoding Transmission Standard), which may become an important industry standard for digital archives. It is also a partner in the DLF Format Registry working group, formed in January 2003 by OCLC, the DLF (Digital Library Federation), and international stakeholders such as JISC, NARA, the National Archives (UK) and JSTOR, which is investigating a registry of data formats to support digital preservation. In addition to these collaborations, OCLC has maintained since 2003 a digital archive. This general purpose OAIS-compliant digital repository accepts materials using OCLCs web-archiving toolset or its batch ingest service. OCLCs digital archive can handle multiple le formats. It links with Connexion (another OCLC service) and offers automatic metadata retrieval with its web harvester service. Its current bit-preservation techniques include circulating backups (onsite and offsite) and media migration, and it is developing processes that will probably combine migration and emulation. (This section is based on Bellinger et al. (2004) and OCLCs web site (www.oclc.org).)
CAMiLEON (www.si.umich.edu/CAMILEON)
CAMiLEON (Creative Archiving at Michigan and Leeds), described in Chapter 7, is noted here because it is a signicant example of an international digital
Regional initiatives and collaborations 169
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
preservation collaboration, in this case a research collaboration. From 1999 to 2003 the University of Michigan (in the US) and the University of Leeds (in the UK) combined forces to develop and evaluate a range of techniques for longterm preservation of digital materials. CAMiLEONs reports and other publications, available on its web site, remain valuable source material. Rushbridge considers its inuence to have been in its high-prole activities, especially the BBC Domesday project (see Chapter 7), and its proof-of-concept of migration on request. CAMiLEON, he suggests, has given leadership, attracted public attention, advanced both theory and practice, and highlighted many of the issues involved in digital preservation today (Rushbridge, 2004a, p.35).
Regional initiatives and collaborations

The term regional is applied here to those digital preservation initiatives or collaborations established to apply to, or be available for, participants from a region greater than a single nation, for example, in the European Union. As with some of the international initiatives and collaborations noted above, the initial intention may have been for it to apply to a narrow geographical region or sectoral limitation, but it may have since become more widely available within a region. The regional service considered here is NEDLIB, and the regional alliances noted are ERPANET and the Digital Recordkeeping Initiative.
Regional services
NEDLIB (www.kb.nl/coop/nedlib)
NEDLIB (the Networked European Deposit Library) was a European Commission-funded research project that was launched at the start of 1998 and ended in January 2001. NEDLIB, a collaborative project of national libraries, a national archive, IT organizations and publishers led by the Koninklijke Bibliotheek (National Library of the Netherlands), explored the technical and managerial issues of building a networked European deposit library. NEDLIBs web site hosts the publications and reports resulting from its activities, which include Jeff Rothenbergs report An Experiment in Using Emulation to Preserve Digital Publications (Rothenberg, 2000). Its outcomes include a proof-of-concept demonstrator of a deposit system for electronic publications, and guidelines for best practice. Its results have been further developed by the Koninklijke Bibliotheek and IBM-Netherlands in implementing the Koninklijke Bibliotheeks electronic deposit system, e-Depot (described below). The signicance of NEDLIB lies in the background work it has done to develop and implement a fully functioning electronic deposit system. (This section is based on Beagrie (2003) and the NEDLIB web site (www.kb.nl/coop/nedlib).)
Regional alliances
ERPANET (www.erpanet.org)
ERPANET (the Electronic Resource Preservation and Access Network) was funded by the European Commission and the Swiss Government from 2001 to 2004 to enhance the preservation of cultural and scientic digital objects through
raising awareness, providing access to experience, sharing policies and strategies, and improving services (ERPANET brochure, 2004). This collaboration is led by HATII (Humanities Advanced Technology and Information Institute) at the University of Glasgow, in partnership with the Schweizerisches Bundesarchiv (Switzerland), ISTBAL at the Universit di Urbino (Italy), and the Nationaal Archief van Nederland. ERPANET has produced a remarkable variety of products and services, aimed at identifying quality sources of information and making their results widely available. They include thematic workshops, training activities, studies of digital preservation activities, assessments of selected publications about digital preservation, an e-repository, tools that provide structured advice to aspects of digital preservation activities, an online discussion forum, and an advisory service. For example (these come from Ross, 2004), it assesses signicant literature on digital curation and preservation selected from a broad range of publications, analyses these, and provides a commentary on them. Its thematic workshops and seminars have covered a wide range of topics, for instance OAIS, digital preservation policies, web archiving, metadata, appraisal of scientic data, le formats for preservation, and persistent identiers. One outcome has been the conclusion that the community at large is hungry for practical case studies and reports of real world experiences, so ERPANET has been conducting case studies in a wide range of sectors, including publishing, pharmaceutical, and government; completed case studies are available on the ERPANET web site. These case studies have led to the awareness that most organizations are hesitant in their digital preservation activities, often because of a lack of awareness, and are waiting for external developments that they can adopt, or off-the-shelf solutions they can implement (Ross, Greenan and McKinney, 2003, p.8). At the heart of ERPANET is its collaborative nature. It depends on synergy with many European and international agencies national libraries, universities, other digital preservation collaborative activities and with professionals (Ross, 2004). At the end of 2004 strategies for continuing ERPANETs key functions beyond the period of its EU funding were being explored. (This section is based on Ross, Greenan and McKinney (2003), Ross (2004) and the ERPANET web site (www.erpanet.org).)
Digital Recordkeeping Initiative

The Digital Recordkeeping Initiative, launched in May 2004, has as its members the government archives of Australia and New Zealand: National Archives of Australia, Public Record Ofce Victoria, State Records New South Wales, Archives New Zealand, Queensland State Archives, Archives Ofce of Tasmania, State Records South Australia, State Records Ofce of Western Australia, Territory Records Ofce of the ACT, and Northern Territory Archives Service. It was formed to develop a collaborative regional Australasian approach to the development, articulation and implementation of a common set of strategies for enabling the making, keeping and using of born digital records. Collaborative work will include determining how to support government agencies to create and manage full and accurate digital records, how to ensure that authentic, reliable and usable digital records are preserved, and how to make these records accessible. A press release about the Digital Recordkeeping Initiative is on the National Archives of Australias web site (www.naa.gov.au/ publications/media_releases/digital.html).
National initiatives and collaborations 171
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
National initiatives and collaborations

The term national refers here to a single country. While the intention may have been to develop a service or alliance within only one country, the availability (as with some of the international and regional initiatives and collaborations noted above) may have since expanded. The national services considered here are the Koninklijke Bibliotheeks e-Depot, the UK National Archives Digital Archive, the Digital Curation Centre, the AHDS, and the UK Web Archiving Consortium. National alliances noted are the Digital Preservation Coalition, the NDIIPP, and the Digital Preservation Testbed.
National services
KBs e-Depot (www.kb.nl/dnp/e-depot/e-depot.html)
The Koninklijke Bibliotheek is responsible for the Dutch legal deposit library, encompassing electronic publications. The Bibliotheek, in partnership with IBM, has developed and implemented a digital archiving system for electronic publications by Dutch publishers. Implemented in December 2002, the e-Depot was fully operational by March 2003. Publishers such as Elsevier Science and Kluwer Academic are participants. By April 2004 more than two million electronic publications had been included in the e-Depot, and it was estimated that by the end of 2004 about 5 terabytes would be held, consisting of 3 terabytes of online journal publications and 2 terabytes of other electronic publications. The e-Depot is based on IBMs Digital Information and Archiving System (DIAS), which is based on the OAIS Reference Model (ISO 14721:2002) and is constructed as much as possible from off-the-shelf components. The e-Depot is designed to meet large-scale and high-quality storage requirements and [to support] digital preservation functionality (The KB e-Depot, 2004). The Library has developed workows for processes such as accepting materials, generating identier numbers, searching and retrieving publications, and for authentication. The e-Depot handles two kinds of digital content: ofine media such as CD-ROMs, and online media such as electronic articles. Ofine media are handled by copying les and the software needed to execute the les to a Reference Workstation. A snapshot of the entire workstation, including its operating system, is then generated into an image which is loaded into the e-Depot. The result is accessed by retrieving the image from the e-Depot and installing it onto a Reference Workstation. Online media are submitted accompanied by publishers metadata, and after checking and validating the content and metadata are linked and the resulting package is then processed. The process is fully automatic. The e-Depots preservation subsystem uses metadata to manage long-term access to the digital materials. Information about le formats and software required to view a stored document is stored. The system indicates when obsolescence is expected so that action can be taken to convert or migrate materials. Currently all e-publishing formats are accepted, although preferred formats may be specied in the future. The Koninklijke Bibliotheek has been active in digital preservation research. One of the projects in which it collaborated was NEDLIB, described above. Other projects include collaboration with IBM, as on the Universal Virtual Computer
(see Chapter 7). The Bibliotheek participates in international working groups such as PREMIS (noted above) and the Digital Repository Certication Task Force, and is active in other digital preservation collaborations. Some of these activities have been adopted elsewhere; for example, Deutsche Bibliotheks KOPAL (Kooperativer Aufbau eines Langzeitarchivs Digitaler Informationen), a long-term archive for digital materials, is based on DIAS. (This section is based on Steenbakkers (2002, 2004a, 2004b), the brochure The KB e-Depot (2004) and the e-Depot web site (www.kb.nl/dnp/e-depot/e-depot.html).)
National Archives (UK) Digital Archive (www.nationalarchives.gov.uk/preservation/digitalarchive)

The National Archives (UK) launched its Digital Archive in April 2003 to store, preserve and provide access to government electronic records. It complies with the UK Governments e-government interoperability requirements and uses open source software, including XML and Java, whenever possible. The tape-based system is designed to handle a wide range of electronic records: documents, e-mails, audio and video les, databases, and web sites. Java-based accession software ensures the creation of metadata and its linking to the appropriate les, carries out quality checks, identies le formats, scans for viruses, and so on. Further technical details are provided on the National Archives web site. The signicance of the Digital Archive has already been acknowledged in winning the 2004 Pilgrim Trust Preservation Award for its innovation in the preservation of digital material. Rushbridge considers that the Digital Archive is likely to have a positive impact among other British institutions. The success of a large-scale, high-prole archive such as this is likely to encourage others to begin their own archiving projects and it may be that such organisations will base their ideas upon those implemented here. This is not without merit: as this essential work needed to be performed for legal reasons the techniques implemented quite rightly prioritise security over innovation (Rushbridge, 2004a, p.34). (This section is based on Brown (2003), Rushbridge (2004a) and the National Archives web site (www.nationalarchives.gov.uk/preservation/ digitalarchive).)
Digital Curation Centre (www.dcc.ac.uk/ )

Funding of 1.3 million a year for three years was announced in February 2004 to establish a Digital Curation Centre in the UK. The funding, provided by JISC and the eScience Core Programme of the UK Government, is to establish a consortium led by the University of Edinburgh in partnership with HATII (at the University of Glasgow), UKOLN (formerly the UK Ofce for Library Networking), and the Council for the Central Laboratory of the Research Councils. Digital curation is a new phrase coined to encompass data archiving, digital preservation, and the active management and appraisal of data over all of its life cycle. The Digital Curation Centre will, according to press releases (5 February and 7 April 2004), provide a national focus for research into curation
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
issues and . . . promote expertise and good practice, both national and international, for the management of all research outputs in digital format. The overriding purpose of the Digital Curation Centre is continuing improvement in the quality of data curation and digital preservation. It is important to note that the Digital Curation Centre will not act as a data repository. Specically, it will lead a research programme addressing digital curation issues, create a network to bring together and support curators, and provide services such as evaluation of tools. A report was commissioned by JISC to establish the current state of, and future requirements for, the curation of research data in the UK (Lord and Macdonald, 2003). The report was in part prompted by the increasing use of large and dynamic datasets in scientic research. Its ndings feed into the Digital Curation Centre and endorse the need for it, and suggest an agenda for research for the Digital Curation Centre. Funding was provided for three years, beginning in April 2004, and the Digital Curation Centre was ofcially launched in November 2004. The activities of the Digital Curation Centre will be watched with considerable interest throughout the world. It has the potential to signicantly advance understanding of digital preservation practice by indicating viable large-scale solutions. (This section is based on Lord and Macdonald (2003), Ross et al. (2004) and the Digital Curation Centres web site (www.dcc.ac.uk).)
AHDS (ahds.ac.uk)
The Arts and Humanities Data Service (AHDS) is a federation of ve data archive services: AHDS Archaeology, AHDS History, AHDS Literature, Language and Linguistics, AHDS Performing, and AHDS Visual Arts. The ve AHDS centres collect, preserve, catalogue, and distribute digital resources which are relevant to their subject areas, facilitate good practice in their creation and use, and offer some user services (ahds.ac.uk/about/index.htm). The AHDS was established in 1996 as an outcome of a JISC feasibility study which recommended and funded a centrally managed distributed service. Its aims were to preserve the rapidly growing unpublished primary digital research materials being generated within the higher education arts and humanities community and beyond (Beagrie, 2001, p.222). It was based on the already existing Oxford Text Archive (established in 1976 at the University of Oxford) and the History Data Service (established in 1993 at the University of Essex), plus three newly-established data archives: the Archaeology Data Service (at the University of York), the Performing Arts Data Service (at the University of Glasgow), and the Visual Arts Data Service (at the Surrey Institute of Art and Design). The holdings of the AHDS include electronic texts, databases, still images, audio, GIS data, geophysics data, and metadata sets. At September 2004 it held approximately one terabyte of data, projected to rise to four terabytes by the end of 2004. A new data repository was planning for a capacity of 10 terabytes. The AHDS sets and guides the overall policy for the management of each of its component services, provides outreach services, sets standards, and encourages best practice. One example is its web-based guides to best practice, such as Digitising History: A Guide to Creating Digital Resources from Historical Documents (Townsend, Chappell and Struijv, 1999).
The AHDS is signicant for digital preservation as an early example of collaboration in data archiving, that is, a centrally managed distributed model. (This section is based on Beagrie (2001) and information from the AHDS web site (ahds.ac.uk)).
UK Web Archiving Consortium (www.webarchive.org.uk)

The UK Web Archiving Consortium was launched in June 2004 as one of the key outcomes of a web archiving feasibility study commissioned in 2002 by JISC and the Wellcome Library. A coalition of the British Library as the lead partner, JISC, the National Archives (UK), the Wellcome Trust, and the national libraries of Scotland and Wales, the Web Archiving Consortium has the goal of investigating archiving solutions for UK web materials. Each institution will focus on a particular subject area, for instance the Wellcome Library on medical sites. The National Library of Australias PANDAS software, further developed to match UK requirements, and software based on the open source web crawler HTTrack will be used. The project is planned to run for two years and has an initial target of 6000 sites. (This section is based on information from the Web Archiving Consortiums web site (www.webarchive.org.uk)).
National alliances
Digital Preservation Coalition (www.dpconline.org)
The idea of a national digital preservation alliance was developed in 1999 at a workshop on digital preservation held at the University of Warwick, which recommended the establishment of a coalition to promote digital preservation activities in the UK. To further the development of a coalition, JISC established a Digital Preservation Focus in 2000 (Beagrie, 2001). The Digital Preservation Coalition (DPC) was established in 2001 with seven founding members. Membership is open to all collectives and non-prot organizations from all sectors in the UK. It had 26 members at January 2004, plus JISC. Its Board of Directors represent the British Library, HEFCE (the Higher Education Funding Council for England), CURL (the Consortium of University Research Libraries), the e-Science Core Programme, National Archives of Scotland, National Archives (UK), OCLC, Public Record Ofce Northern Ireland, Resource (the Council for Museums, Archives and Libraries), and the University of London Computer Centre (DPC Annual Report, 200203, p.1). The DPCs aim is to secure the preservation of digital resources in the UK and to work with others internationally to secure our global digital memory and knowledge base. To do this it carries out a wide range of activities: producing, providing and disseminating information on current research and practice and building expertise among its members getting digital preservation on the agenda of key stakeholders securing funding providing a common forum for the development and co-ordination of digital preservation strategies in the UK promoting standards, services, and technology forging strategic alliances with relevant agencies nationally and internationally
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
working collaboratively together and with industry and research organisations, to address shared challenges in digital preservation (DPC Annual Report, 200304, p.1). Its activities in 20022004 included an advocacy campaign (for example, publishing articles and issuing news items), representation and advocacy (such as making presentations at conferences), dissemination and current awareness (through its web site, collaboration with the National Library of Australia to produce the e-newsletter Whats New in Digital Preservation, and e-mail discussion lists), forums and training workshops, a survey of industry vendors, work on a National Needs Survey on Digital Preservation, its Technology Watch, and surveying DPC members as part of a UK-wide needs assessment exercise (DPC Annual Report, 200203, 200304). The DPC co-sponsors a Digital Preservation Award as part of the Pilgrim Trust Conservation Awards. The inaugural award, in 2004, went to National Archives (UK) for their Digital Archive; other shortlisted projects were the CAMiLEON Project, the JISC Continuing Access and Digital Preservation Strategy, the National Library of New Zealands Preservation Metadata Extraction Tool, and the Wellcome Library/JISC Web Archiving Project. The DPCs collaborative activities are not limited to the UK: for example, it has signed Memoranda of Understanding with the National Library of Australia and the Library of Congresss NDIIPP (Jones, 2004). The signicance of the DPC for digital preservation lies in its active encouragement of digital preservation through its lobbying and promotional activities. It is being emulated in other countries; for example, the German Network of Expertise in Long-Term Storage of Digital Resources (NESTOR), led by Deutsche Bibliothek, states that its long-term goal is to develop a permanent distributed infrastructure for long-term preservation and long-term accessibility of digital resources in Germany comparable to e.g. The Digital Preservation Coalition in the UK (www.langzeitarchivierung.de). (This section is based on Beagrie (2001), the DPC Annual Reports for 200203 and 200304, Jones (2004), Simpson (2004b) and information from the Digital Preservation Coalitions web site (www.dpconline.org)).
NDIIPP (www.digitalpreservation.gov)
The US-based National Digital Information Infrastructure and Preservation Program (NDIIPP) is led by the Library of Congress. Federal legislation established this programme in December 2000 with US$100 million in funding, US$25 million being made available to the programme and a further US$75 million to be made available if it matched the amount from non-federal government sources. NDIIPPs goals are to develop a national digital collection and preservation strategy, to work with other stakeholders to establish partnerships and form networks, to help identify and preserve at-risk digital content, and to support the development of tools, models, and methods for digital preservation. Collaboration lies at the heart of the NDIIPP, and is in fact mandated by the legislation that created it, with collaboration expected between the Library of Congress, National Archives and Records Administration, National Library of Medicine, National Agricultural Library, National Institute of Standards and Technology, and other federal agencies as well as non-federal organizations and institutions. Since its establishment the NDIIPP has sought participation from a wide range of organizations (Hodge and Frangakis, 2004, p.62).
Principal activities of the NDIIPP to date have included convening stakeholder meetings in 2001, commissioning environmental scans, and presenting a plan to Congress in 2003. The NDIIPP commissioned a report on international digital preservation activities for its own information, the initiatives surveyed being selected for their relevance and interest to NDIIPP (Beagrie, 2003). In mid-2004 Version 2.0 of the Technical Architecture for NDIIPP was released for review and comment, and NDIIPP entered into a partnership with the National Science Foundation to fund research programmes in digital preservation. In June 2004 the Library of Congress announced that it would collaborate with Old Dominion, Johns Hopkins, Stanford and Harvard Universities in an Archive Ingest and Handling Test that would investigate strategies for the ingest and preservation of digital archives. Each partner would work with the same collection of digital materials, 12 gigabytes of materials with a wide range of le formats relating to the events of 11 September 2001 made available by George Mason University. The outcomes of this research will include better identication of possible architectures, better costing data, and clearer identication, documentation and dissemination of working methods for preserving digital materials. In September 2004 the NDIIPP awarded US$15 million to eight US consortia for three-year projects to identify, collect and preserve digital materials within a nationwide digital preservation infrastructure. To illustrate the extent of these projects, one of these consortia is led by the California Digital Library (University of California) with New York University Libraries, University of North Texas Libraries, and the Texas Center for Digital Knowledge as partners, and the San Diego Supercomputer Center, Stanford University Computer Science Department, and Sun Microsystems as collaborators. It will develop web archiving tools to capture and preserve web-based collections of government and political information. Another programme is aimed at establishing procedures, structures and national standards for preserving public television programmes produced in digital formats, and a third will develop criteria for determining which digital materials to capture and preserve. The NDIIPP is in 2005 exploring collaborative arrangements with US state and territory governments to develop strategies for preserving information produced by state and local government jurisdictions in digital form. After ve years, the NDIIPP considers its achievements to include developing a network of 50 to 75 partners, establishing a large archive of at-risk content, and making recommendations to the US Congress about the long-term governance of a national digital preservation programme. The NDIIPP is being keenly observed by digital preservation interest groups throughout the world. With such large resources at its command it stands a good chance of providing tools that will revolutionize how digital preservation is carried out. (This section is based principally on information from the NDIIPPs web site (www. digitalpreservation.gov).)
Digital Preservation Testbed (www.digitaleduurzaamheid.nl/home.cfm)

The Digital Preservation Testbed (Testbed Digitale Bewaring) was established by the Dutch Government as a three-year project to run from October 2000 to September 2003. It was part of a wider initiative, the Digitale Duurzaamheid Project, which established a Taskforce Digitale Duurzaamheid (Digital Longevity Taskforce) to coordinate and share the expertise of government organizations in
Sectoral initiatives and collaborations 177
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
the Netherlands about digital information management. The Digital Preservation Testbed project was launched as part of this initiative. Based at the Nationaal Archief of the Netherlands, it carried out research into the digital preservation techniques of migration, emulation, and XML, applying them to a range of digital materials such as text documents, spreadsheets, e-mail messages, and databases. Each technique was assessed for its technical effectiveness as a longterm preservation method, as well as for cost factors and other aspects. Reports of the project are available on the web. (This section is based on Beagrie (2003, p.28) and information available on the web site of the Nationaal Archief of the Netherlands (www.digitaleduurzaamheid.nl/home.cfm)).
Sectoral initiatives and collaborations

The term sectoral refers to a sector such as a discipline, interest group, or industry. Examples of sectors included in the examples below include the higher and further education sector in the UK, and a discipline-based service aimed at linguists and ethnomusicologists. Case studies of other sectors in Europe are available in the ERPANET case studies, for example the pharmaceutical, publishing and government sectors. (For more information see the section on ERPANET above, the ERPANET web site, and Ross, Greenan and McKinney (2003)). The sectoral services considered here are PARADISEC and CEDARS, and the sectoral alliance noted is JISC.
Sectoral services
PARADISEC (paradisec.org.au)
PARADISEC (Pacic and Regional Archive for Digital Sources in Endangered Cultures) recognizes Australias location in a region of linguistic and cultural diversity, and addresses concern about the likely loss of languages spoken in the Pacic region (broadly dened to include Oceania and East and Southeast Asia) from 2000 to only a few hundred in the next hundred years. It is creating an archive of sound recordings of languages of the Pacic Islands, most of which are endangered and sparsely documented. PARADISECs focus is on preservation of the content of eld tapes from the 1950s and 1960s, most of which are on magnetic tape that is now deteriorating. It therefore provides a facility for accessioning, cataloguing, digitizing and preserving this material. PARADISEC, which received grant funding in 2001, is a collaboration of the Australian National University and the Universities of Sydney, Melbourne, and New England. It is managed from the University of Melbourne, and its facilities are at the University of Sydney. It digitizes the audio data at 24-bit, 96kHz in BWF (Broadcast Wave Form) using a Quadriga system and AudioCube workstation, and adds metadata that conforms to Dublin Core and OLAC (Open Languages Archives Community) standards. Backup copies are stored in Canberra. Other activities include hosting workshops, and providing services on a cost recovery basis. PARADISEC is an example of the value of collaboration to achieve large-scale digital preservation outcomes. (This section is based on Pawley (2003), a PARADISEC information brochure of August 2004, and the PARADISEC web site (paradisec.org.au)).
CEDARS (www.leeds.ac.uk/cedars)
CEDARS (CURL Examplars in Digital Archives) was a collaborative research project funded by JISC that ran from 1998 to 2002 and hosted at the Universities of Leeds, Oxford and Cambridge. It is included here because it was an important test-bed for digital preservation within research libraries in the higher education sector (Beagrie, 2001, p.219). Its aim was to provide guidance to others in the sector about best practice for digital preservation. Among its outcomes are a preservation metadata schema and an archiving demonstrator project based on OAIS. Reports generated by the project are available from its web site. Some of the activities of CEDARS are noted in Chapter 5. The signicance of CEDARS for digital preservation lies in its reports (available on the CEDARS web site) and its development of a prototype distributed archiving system. (This section is based on Beagrie (2001) and the CEDARS site (www.leeds.ac.uk/cedars)).
Sectoral alliances
JISC (www.jisc.ac.uk)
JISC (the UK Joint Information Systems Committee) represents over 700 institutions in the UK higher and further education sector. Its mission is to promote the use of ICT in that sector. It has energetically pursued digital preservation activities through collaborations such as the AHDS, CEDARS, and CAMiLEON (all of which are referred to elsewhere in this chapter). Beagries description of JISCs Continuing Access and Digital Preservation Strategy (one of the nalists in the Pilgrim Trusts inaugural Digital Preservation Award), notes that JISC has three major objectives in its digital preservation activities: establish best practice and guidelines for digital preservation and to disseminate them here a major outcome has been the publication of Preservation Management of Digital Materials: A Handbook (Jones and Beagrie, 2001) generate support and collaborative funding from, and promote interworking with, agencies worldwide the DPC is the principal outcome of this objective develop a long-term digital preservation strategy for digital materials of relevance to Higher and Further Education in the UK the major outcome is JISCs Continuing Access and Digital Preservation Strategy and implementation plan (Beagrie, 2004). JISCs Continuing Access and Digital Preservation Strategy is wide-ranging. It includes the provision of a wide range of resources on its web site, for example internet resources, e-journals, e-prints, feasibility studies and risk assessments that recommend actions for specic categories of digital material of interest to JISC members. It lobbied successfully for a Digital Curation Centre (described above). Beagrie notes other JISC achievements in progressing digital preservation in the UK: Advocacy and funding, for example over 6 million in funding has been secured over the next three years, representing a substantial increase over previous funding Infrastructure, most notably the Digital Curation Centre
Conclusion
179
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Supporting a life-cycle approach to the management of digital resources Scoping studies: an example is the 2003 Feasibility and Requirements Study on Preservation of E-Prints: Report Commissioned by the Joint Information Systems Committee (JISC) (James et al., 2003). Partnerships, for example with DPC, and the UK Web Archiving Consortium (Beagrie, 2004). JISC is signicant because it has been an important catalyst for digital preservation in the UK, not only in the higher and further education sector, but also more widely in the UK and internationally. (This section is based on Beagrie (2004) and on the JISC web site (www.jisc.ac.uk).)
Conclusion
The intention in describing this selection of digital preservation programmes and initiatives is to illustrate the range and nature of current digital preservation activities, and to emphasize their increasingly collaborative nature. They may, perhaps, prompt the reader to reect on digital preservation activities, and in doing so, to derive some value from them. Many other activities could have been included, such as those listed in the UNESCO Guidelines (UNESCO, 2003, pp.6869): The Austrian On-Line Archive (AOLA), an archive of snapshots of Austrian web space (www.ifs.tuwien.ac.at/~aola/) The Academic Research in the Netherlands Online (ARNO) which links the document servers of three Dutch universities to keep their digital academic output accessible (www.uba.uva.nl/en/projects/arno/) The Australian Digital Theses Project, a distributed database of digital versions of theses produced by postgraduate research students at the participating institutions (adt.caul.edu.au/) The Digital Image Archive of Medieval Music (DIAMM), a collaboration between the University of Oxford, the University of London and AHDS, a permanent electronic archive of European medieval polyphonic music (www.diamm.ac.uk/) the European Visual Archive (EVA), providing easy and preserved access to the integrated collections and information held in European archives (www.eva-eu.org/). A useful starting point to locate other digital preservation services and initiatives is the erpaDirectory, available online from the ERPANET web site (www. erpanet.org). Descriptions of other digital preservation services and initiatives are found in the case studies in the Appendix of this book, which include descriptions of activities at the National Library of Australia (PANDORA and PADI), VERS at the Public Record Ofce Victoria, the National Archives of Australia, and the Department of Family and Community Services. The reader should be aware that these descriptions are of activities up to the end of 2004. Given the rapid rate of progress in the eld of digital preservation, frequent reference to the web sites of these programmes and initiatives, and to sources such as the PADI portal, is recommended to update knowledge of each of them.
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Chapter 10
Challenges for the Future of Digital Preservation
Introduction
The lack of awareness of digital preservation issues by stakeholders, the lack of the necessary skill sets to preserve digital materials, the lack of agreed international approaches, a shortage of practical models on which to base preservation practice, and a lack of funding on an ongoing basis to address digital preservation issues, all contribute to the problem (Beagrie, 2003, p.6) We have more technology than we can deal with . . . we lack stable structures (nancial, expertise, organizational) to build on to (Geser and Pereira, 2004, p.59) This chapter attempts to provide some pointers to those challenges that are likely to be the focus of digital preservation activities in the future. Predicting futures is, of course, always a risky business; in this attempt to do so the views of many experts are drawn upon to derive a consensus about what the challenges will be. A brief reiteration of the key threats to digital continuity is worthwhile to remind us of the issues. The list of obstacles to preserving digital heritage in CASL (Council of Australian State Libraries) member libraries is one brief statement of the threats; it includes uncertainty about selection for preservation, archiving and storage, authenticity, intellectual property rights/digital rights management, technological obsolescence, costs, and roles and responsibilities (Council of Australian State Libraries, 2002). A fuller list comes from Jones and Beagrie (2001, p.2734): rapid obsolescence resulting in large part from a market-driven ethos the creators of digital materials and other stakeholders may lose interest the lack of awareness of digital preservation issues by stakeholders
182 Challenges for the Future of Digital Preservation
the lack of the necessary skill sets to preserve digital materials the lack of agreed international approaches a shortage of practical models on which to base preservation practice a lack of funding on an ongoing basis to address digital preservation issues traditional organizational structures that do not readily accommodate digital preservation activities lack of agreement about which institutions should preserve digital materials lack of applicable selection principles legal issues, such as intellectual property rights and legal deposit.
What have we learned so far?

How far have we gone towards countering the threats to digital continuity? What do we now know about digital preservation? Some of our knowledge to date is summed up in three aphorisms developed from the experiences of participants in Canadian Conservation Institute workshops: something is better than nothing; move it or lose it; dont throw out the original (Strang, 2003). Webbs six points also summarize a signicant body of experience: selection is necessary benign neglect does not work someone needs to accept responsibility working together is effective and increasing we already know much, including much about what we need to do action is possible now, even if we do not have all of the answers (Webb, 2004, p.50 (paraphrased)).
A third statement of emerging Practices the authors suggests that it may be premature to identify Best Practices) is provided in the responses to the PREMIS survey carried out in 20032004. It includes: Use of the OAIS model as a framework and starting point for designing the preservation repository . . . Maintenance of multiple versions (originals and at least some normalized or migrated versions) . . . Choice of multiple strategies for digital preservation (OCLC/RLG PREMIS Working Group, 2004, p.54). These, however, are only summaries; closer examination of the challenges is worthwhile to reinforce how much we still need to do and to delineate more precisely those challenges. The challenges are not just technical; in fact, there is a school of thought that we have most of the technical solutions that we need at hand. The challenges also include key social, political, and economic issues . . . including the need to develop a will to preserve and provide permanent access (Hodge and Frangakis, 2004, p.3). The 14 challenges here do not constitute a comprehensive list, but are those about which there is a high level of consensus. 1) Developing standards for digital preservation and encouraging their use. The development and implementation of standards (discussed in Chapters 7 and 8)
What have we learned so far? 183
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
are a key aspect of most digital preservation activities. The benets of using standards include the opportunity for increased interoperability, fewer problems and greater economies associated with limiting the number of le formats handled, and facilitation of collaboration. The rapid rate of change associated with technologies, for instance, the rapid obsolescence of le formats as new formats replace them, means that monitoring and maintenance of standards needs to be constant and ongoing. 2) Inuencing data creation where possible. Encouraging the application of standards such as metadata standards, persistent identiers and non-proprietary le formats that will prolong the effective life of the available means of access and reduce the range of unknown problems that must be managed (UNESCO, 2003, p.37) is a challenge that must be addressed to ensure effective digital preservation. This is related to the need to engage all stakeholders in the process, noted in Chapter 2 and also below. 3) Increasing collaboration and engaging stakeholders in digital preservation. Chapter 9 noted in some detail the value of collaboration and Chapter 2 the value of engaging stakeholders in digital preservation activities. Working collaboratively is often a cost effective way to build preservation programmes with wide coverage, mutual support and the required expertise (UNESCO, 2003, p.25). The greatest challenges are in identifying stakeholders and potential partners and engaging them in collaborative activities; other challenges lie in managing collaborative activities so that they are effective and benet all participants. 4) Developing policy about digital preservation. Clearly articulated policies, such as a written policy framework with digital longevity as an explicit goal (B. Smith, 2002, p.136), although not noted in detail in preceding chapters, is considered an essential requirement for effective digital preservation. One aspect of policy, and perhaps the one where the greatest challenges lie, is the decision about who will take responsibility, critical because digital materials require constant and ongoing attention to ensure their preservation. Acceptance of responsibility should be explicitly and responsibly declared (UNESCO, 2003, p.23). 5) Ensuring that legal rights do not impede digital preservation. A major challenge (noted in Chapter 4) results from the conict between ownership and control of intellectual property, on the one hand, and the needs of cultural heritage institutions to collect and copy for preservation purposes, on the other. Ongoing efforts are required in order to change legislation to ensure adequate rights to collect and copy digital materials for preservation purposes. 6) Determining which digital materials, and what attributes of them, it is crucial to maintain access to. Although some maintain that we will soon be at the stage where it is technically feasible to collect all digital materials, most commentators appreciate that it is not practical to collect everything, so selection of digital materials is required (noted in Chapter 4). In addition, decisions about the attributes of digital materials that we should preserve are required (noted in Chapter 5). Clearly articulated selection policies and guidelines must be developed and a better understanding of the attributes of digital materials that affect their authenticity must be reached. 7) Integrating digital preservation into mainstream operations. It is now recognized that the viability of ongoing digital preservation programmes is compromised if they depend on short-term project funding, and that they need to be fully incorporated into the mainstream activities of institutions. The challenge is to
build sufcient awareness of this so that digital preservation is appropriately funded and resourced as part of mainstream activities. Aspects of this challenge are discussed further later in this chapter. 8) Taking action now, even if on a small scale. There are, as yet, no widely agreedupon techniques for digital preservation, although (as noted in Chapters 6, 7 and 8) the seeds of agreement are beginning to germinate. Current advice from heritage institutions with signicant digital preservation experience is to take active steps now, even small ones, which will preserve access for the manageable future, while also planning for whatever long-term approaches appear to be the most practical (UNESCO, 2003, p.124). Waiting until reliable techniques are available will almost certainly result in the loss of material. This challenge is not explicitly noted in preceding chapters; Chapter 5 of the UNESCO Guidelines (2003) stresses its importance. The challenge is to promote greater awareness of this and provide appropriate guidelines about how to do it. There is more detail about aspects of this challenge later in this chapter. 9) Maintaining access to the preserved digital materials. Perhaps the greatest challenge faced in digital preservation is to maintain access the ability to present the essential elements of authentic digital materials (UNESCO, 2003, p.22) to preserved digital materials. The numerous challenges in achieving this are noted throughout this book, and the strategies and techniques available to us are noted in Chapters 6, 7, and 8. 10) Maintaining the authenticity and integrity of the preserved digital materials. This challenge is noted in detail in Chapter 5. Although authenticity is not a primary requirement for all digital materials, for many categories (for example, where the digital materials are used for evidential purposes) it is essential. One challenge, as noted above, is to determine the attributes that determine authenticity for different categories of materials; another is to maintain those attributes into the future. 11) Developing better costing data. There is a lack of reliable data about the costs of digital preservation, in part because the techniques are rapidly evolving as our knowledge of the requirements matures, in part because of the massive quantities of digital data, and in part because of the ongoing costs (the digital mortgage referred to in Chapter 4). Better costing data is needed immediately; lacking it, actions such as securing funding and determining responsibilities for digital preservation will be impeded. 12) Securing funding for digital preservation. As noted above, an emphasis on short-term project funding is not conducive to long-term viability of digital preservation. Securing long-term funding requires that funding agencies and political masters must be convinced of the need for the preservation of digital information. It may, in part, be addressed by success in addressing one of the challenges mentioned above integrating digital preservation into mainstream operations 13) Increasing awareness about digital preservation. Digital preservation will be incorporated into mainstream activities, will attract secure funding, will attract appropriate resources such as skilled personnel, and in other respects will achieve maturity only when awareness of it is heightened at all levels, from the political to the individual. 14) Developing the skills required for digital preservation. Skilled personnel with appropriate knowledge of digital preservation requirements are in short supply around the world (noted briey in Chapter 1). The challenge is to identify more
Four major challenges 185
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
precisely the skill sets that are required, and to encourage their acquisition. This challenge is discussed in more detail later in this chapter.
Four major challenges

Of these 14 challenges, four emerged as the most intractable during interviews with Australian digital preservation experts in 2004: challenge 1: managing digital preservation, especially the importance of integrating digital preservation into mainstream operations challenge 2: funding digital preservation, especially without available costing data for digital preservation activities challenge 3: peopling digital preservation, especially nding the skills required in the future and ensuring that they remain available challenge 4: making digital preservation t, especially achieving scalability of practice to both large systems and the smallest institutions. The views and comments gathered during the interviews form the basis of the following discussion. Unattributed quotations in the rest of this chapter come from these interviews.
Challenge 1: managing digital preservation

One frequently voiced issue is that digital preservation has not become a normal part of mainstream practice in most institutions. Much of the development to date has been carried out as special projects. The project approach has some merit as a response to the magnitude of the problem, where project-based approaches that bring stakeholders together to collaborate are desirable and effective. It also has merit as a way of securing funding for scoping studies, or of exploring a technique, or of testing the waters. However, project-based funding is by denition short-term and engenders short-term ways of thinking. Project-based digital preservation activities get in the way of an appropriate recognition that digital preservation is here to stay and can only be addressed if all institutions and stakeholders play a part. Beagrie states that ultimately, digital preservation will be successful when it can be seen not as a stand alone institutional activity but as an activity embedded in how institutions manage and approach digital information and resources on an ongoing basis . . . It remains a simple objective, yet one immensely challenging to achieve (Beagrie, 2004). The project-based approach has achieved some notable successes, but it will not ensure sustainability in the medium or long term. The challenge is to integrate digital preservation fully into normal operations of libraries, archives, museums and other institutions with responsibility for digital sustainability. As one Australian digital preservation specialist put it, the challenge weve got . . . around digital preservation . . . [is] that it just becomes the way we do things. An example of the dangers of short-term thinking was provided from a museum context:
the emphasis is on not only short-term but immediate access within an exhibition environment, and as soon as thats over, they move onto the next exhibition . . . Gather[ing] documentation . . . was seen as being totally irrelevant . . . it was something that was done for the short-term, but not seen as relevant for the long-term, whereas in fact it is vitally relevant for the long-term, but there was no interest in the long-term, because the interest was only for as long as the exhibition. And this was just a different set of values . . . a curatorial value . . . that was the space they operated in. And that tended to suck the resources away from any long-term thinking. Another person interviewed described the issue in the library context. Because digital materials have not been incorporated into the normal processes, they are treated as a different acquisition process, or as a different this or a different that. One of the difculties arises with the interactions across different areas, because when you talk about a process that produces a digital outcome . . . typically they go across all the more standard business units within the library, and were often quite split up between technical services and reader services and systems services . . . Whereas a new electronic service cuts across all three of them, and in fact has no business owner at all. And again, I guess thats one of the issues about digital preservation. Who is the business owner? Who owns that within the library environment? In our case it has come out of being owned by the systems area, which is what Im responsible for, but ultimately it needs to be owned by the normal processes and not be seen just as an IT problem, an IT issue. The principal reason for this concern is that sustaining digital materials over time requires ongoing, unbroken activity, expressed variously by Australian digital preservation specialists in this way: preservation is not something that you ever achieve. Its just a process youre in the middle of, or at the start of, but never at the end of; one of the things that is key . . . is recognition that it is an active process, because . . . you do have to think about . . . active management of that object through its life, through time; we need to move away from the sort of proselytizing kind of mode into just making it mainstream. How do we make the change from project-based, short-term support for digital preservation to integration into mainstream activities? What changes will be required? One action suggested by an Australian digital preservation specialist to break down the existing structures in his library was to create a new position at senior level called an E-services Integration Manager, whose job is to work across different areas and ensure . . . that the disconnects dont happen, but the connects are in place between this business unit and that business unit. So that the ultimate electronic service that emerges is a proper continuum of the working out internal inputs and outputs. He also acknowledged that its not as simple as getting . . . new staff dedicated to doing this; it has to be built into the workows of current staff. Some institutions have already successfully integrated digital preservation into their normal operations, or are well on the way to doing it. One Australian example is the National Archives of Australia. Its approach was described in this way:
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
We say: Well the preservation programme is concerned to preserve all formats of records one of the formats is digital preservation. It requires different skill sets and so on, but its about preservation. Its not about digital . . . just as AV preservation is about preservation of the record there are special skill sets for people needed in that too, but its still fundamentally about preservation. Another is the National Library of Australia, who has been able to institutionalize [digital preservation] or operationalize it. This has proved successful because digital preservation is moved out of the realms therefore of just being fashionable and is therefore, less vulnerable to changes in funding. Two approaches to encourage integration are gaining popularity. One is a risk management approach, increasingly being applied to digital preservation activities. This was described as both identifying potential risks more comprehensively and taking action within a whole range of different ways to manage those risks. It also does mean accepting some risk as inevitable. Another approach is to use business planning methodologies (see Bishoff and Allen, 2004).
Challenge 2: funding digital preservation

The UNESCO Guidelines note that two aspects should characterize reliable digital preservation programmes. One is organisational viability which is based on an ongoing mandate and appropriate resources and infrastructure. The other is nancial sustainability: a likely prospect of the organisation being able to continue to provide the required resources well into the future, with a sustainable business model to support its digital preservation mandate (UNESCO, 2003, pp.4344). As noted in the discussion above, digital preservation is vulnerable unless it becomes part of core business activities. This means that its funding must become a core business expense and, in the library context, that preservation needs to be viewed as an activity just as essential as is cataloging (A. Smith, 2002, p.5). Funding should be sustained to pay the digital mortgage. Adequate sustained funding is critical; in fact, one interviewee suggested that it was the issue for digital preservation, that the technical issues such as the intellectual challenge in working out what the essence of those new data formats are as they get invented and developed and making sure we capture those . . . really resolve into resourcing issues, because the expertise to do that stuff, to do the technical work is around, we just have to be able to afford to buy it. There is a reliance on an industrial infrastructure which becomes ever more complex; high-level computer systems are needed and theyre not cheap and theyre not cheap to run. Furthermore youve got to replace everything every ve years. Repositories typically continue to grow in size, so the funding requirements also continue to grow. Even for one well-established Australian digital preservation programme, one of the hardest things has been to continue the funding. Its been an ongoing challenge to continue funding the program. This programme has succeeded in securing funding, in part because we implemented practical projects as weve moved forward.
Short-term thinking based primarily on business need was described by one Australian digital preservation expert. For his organization, preservation is only seen by the business as being valuable if its continuing to meet a business need. If he wished to fund digital preservation activities, the only way Im going to be able to do that is to link it into the drivers in that particular sector. Where the business reasons for keeping materials disappears, the default will be that we will dispose of them. Or we will maybe start selectively disposing of it, so well start prioritizing those that were going to preserve into the future and well be pretty ruthless about doing that. One problem that continues to get in the way of securing ongoing funding is our lack of concrete knowledge of how much it will cost. Earlier projects such as CEDARS attempted to develop rmer costing data as part of their research activities, but despite their efforts the outcomes are still not well dened. Research is ongoing, with initiatives like the NDIIPP in the US including the development of better costing data as part of the research it promotes. Schoneld et al. (2004) report on their investigations into the ongoing costs to libraries of maintaining periodicals in print and digital forms. Sanett and Cloonan point out that the major factor that inuences decisions in many institutions is how much it will cost, and they urge the development of better cost modeling (Cloonan and Sanett, 2000, p.210). Sanett notes elsewhere that a cost model makes intelligent planning possible (Sanett, 2003). One attempt to cost repository storage at OCLC notes a 1996 prediction by Lesk that the costs of the digital and traditional library operations [would] cross over in about ve years, and that electronic storage would offer a major cost advantage within ten (Chapman, 2003, p.8). The OCLC experience concluded that this cross-over point had not been reached for one type of digital material, compressed 600 dpi images, but had been for ASCII les. The study concluded that there are many variables to the equation of managed storage costs, such as formats, numbers of items, number of versions, and number of collections, to allow rm costings at present (Chapman, 2003, p.13). James et al. (2003) have also researched costs of repository storage, in this case cost models for preserving e-prints in institutional repositories. Lavoie has developed a model that describes the economic issues of funding digital preservation in terms of incentives, which need to be sufcient so that stakeholders will participate. The key stakeholders are the owners of the intellectual property rights, the archive that provides the services, and the entity that benets from the long-term preservation of the materials. Lavoie characterizes the incentives as perceived motivation sufcient to 1) induce a party to recognize a need to take action to secure the long-term viability of digital materials in which they are a stakeholder, and 2) induce a party to develop and implement technologies aimed at ensuring the long-term viability of digital materials (Lavoie, 2003, p.ii). He notes that the absence of empirical data about costing has required that values for many of the variables be estimated (2003, p.7). He examines ve organizational models and concludes by providing an agenda for further research: A thorough understanding of the nature and extent of the incentives to preserve is fundamental to more focused research into the economics
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
of digital preservation. Future research activity should include the accumulation and synthesis of digital preservation case studies; the development of appropriate policies for enhancing incentives based on the characteristics of the underlying organizational model; characterizing and analysing the structure of aftermarkets associated with digital preservation services; and devising sustainable pricing strategies for digital preservation services (Lavoie, 2003, p.iii). Lavoie has since expanded his view of the economic issues to encompass the responsibilities of stakeholders, strategies for organizing preservation resources most efciently (Lavoie, 2004), and public good in relation to responsibilities and costings for digital preservation (Lavoie and Dempsey, 2004). There are some hopeful signs. In Australia, changes to public administration policy and accounting practice at federal and state government levels has had the effect of providing more funding for digital preservation. The move to accrual accounting has valued collections as assets and enabled funding to be provided for asset maintenance and replacement (Stuckey, 2003). In the case of the National Archives of Australia this has meant that AUD$15 million in funding was available from the federal government that can only be spent on activities that make the collection last longer, including digital preservation.
Challenge 3: peopling digital preservation

Commentators regularly lament the lack of people with the expertise needed to further the digital preservation agenda. This is not a new situation. Lack of staff expertise is a common problem both in institutions with digital preservation responsibilities and in institutions that have not yet assumed responsibility for digital materials, noted Hedstrom and Montgomery in 1999, adding that those with the required information technology skills typically did not have an understanding of long-term preservation The lack of expertise in digital preservation appears to be a signicant obstacle to developing digital preservation programs (Hedstrom and Montgomery, 1999, pp.1618). More recently the International Council on Archives noted the lack of adequate training of, and human resources development for, records personnel and the impediment this was to the efforts of archivists to protect archives as evidence and to preserve an authentic digital heritage (Millar, 2004, p.5). What new knowledge, what new skill sets are needed? Chapter 1 noted that new skills would be required to implement new policies and develop new procedures. Jones and Beagrie describe the digital preservation work environment and identify some of its requirements. The general work environment is one of rapid and constant change, blurring of boundaries of responsibilities, and increased weight given to collaboration and accountability. For digital preservation activities specically, formal training opportunities are rare, so that much is learned on the job, and informal contacts with others working in the eld are valuable (Jones and Beagrie, 2001, p.54). The Australian digital preservation specialists interviewed provided some informative perspectives. While most of them considered that a new kind of digital preservation specialist was required, one counter view was expressed. One archivist could not see that there is a new profession in digital preservation because . . . digital preservation is just a process for preserving digital
records and its an IT process, so anyone with IT skills can do it. He made an exception for some research and development work in digital preservation, which would require new skills. These Australian specialists considered that general professional skills were required to a high level. One indicated that a preservation specialist in any medium (digital being no exception) needs to rst understand the purpose of the whole activity. The need for thinking about [digital preservation] holistically as opposed to in a narrow way was also noted by another interviewee. Others described the general characteristics and skills that were required: for example, the most important characteristics of people who work in this area are things such as [an understanding that] . . . theres no such thing as a wrong answer; the open mind, the inquiring mind was needed. Specic skills required were incredibly good project management skills, diplomacy (youve got to have an ability to work across complex organizations without treading on anybodys toes), and communication skills (the ability to write a sentence which more than two people agree with). People able to take a strategic perspective, and who could put a business case were considered necessary. The major qualication required, one interviewee suggested, was managerial skills, because its a process, so you have to manage the IT specialists who are delivering it. Technical skills to a higher degree than previously required were considered necessary. At a broader technical level, digital preservation specialists are obviously going to need to know a lot about metadata and databases and about the specic media that they are working with. Additionally there is a need for a layer of specialist information about how metadata works in the preservation context and how digital preservation systems work . . . there is a specialty there and it probably needs to be available as an add-on to things that are more content-related. All interviewees agreed that new combinations of skills were required, adding digital knowledge to analogue knowledge. The IT skills needed to be at the level where youve got the literacy to be able to get down into that bitstream and extract things and pull things out of corrupted disks, and all of that all that technical nous, and, in addition to this computer literacy, digital preservation specialists need to have skills in change management and an ability to liaise with owners and creators of digital collections, to provide ongoing guidance and write comprehensible guidelines for depositors, and to ensure rigorous documentation and ongoing maintenance procedures are undertaken. Working as a team will also be a pretty important as theyll certainly need to do that! The important thing to note is that it wont just be technical skills which are required. This new curriculum should be based on a common body of knowledge which can form the basis of a set of competencies and should be available in a variety of modes, such as learning by doing, secondments, workshops which include practical case studies, as continuous professional development courses or formal training courses at undergraduate and postgraduate levels. The people who would become digital preservation specialists could potentially come from any background. However, the experience of one interviewee was that the people we are nding really good are actually coming out of history and the humanities. He continued that this was not to say that IT people werent appropriate;
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
however, his experience was that they will often come with limitations set by the technology . . . theyll be thinking about the technology underneath [rather than] thinking much more broadly about why we are doing this and what this could mean in the future. Some formal training programmes are available. HATII (the Humanities Advanced Technology and Information Institute) at the University of Glasgow offers a full award, the Master of Philosophy in Information Management and Preservation, and other schools of information management or librarianship offer individual subjects. Less formal programmes, such as online self-study programmes, are being developed. One example is the Cornell University Librarys online tutorial Digital Preservation Management: Implementing Short-term Strategies for Long-term Problems (Kenney et al., 2003). Digital preservation training is recognized by the DPC as a major issue: it remains difcult to recruit individuals with appropriate skills and the need for such skills is growing (Cornwell Management Consultants, 2004, p.1). The DPC, therefore, carried out a training needs analysis in 2004, which recommended that courses be developed and offered (Cornwell Management Consultants, 2004); in conjunction with JISC and the British Library, the DPC is developing a modular training programme (JISC, 2004). The challenge of meeting the pressing need for education and training in new digital archiving methods, tools, and technologies (Workshop on Research Challenges in Digital Archiving and Long-term Preservation, 2003, p.xix), is large and will take time to fully address.
Challenge 4: making digital preservation t

Anne Kenney, in an unpublished address to the UNESCO Regional Consultation on the Preservation of Digital Heritage meeting, Canberra, Australia, 46 November 2002, noted that digital preservation activities can be characterized as being too high (conceptual rather than pragmatic), too low (narrowly dened), too little (failing to acknowledge the problem), or too late (occurring reactively rather than anticipating the problem). If digital preservation activities are to be fully integrated into standard practice they need to be applicable to all sizes and shapes of institutions. People who work in these institutions must be aware of the need for digital preservation and must have ready access to advice that is applicable to their institutions. Discussion about digital preservation has too frequently been couched in terms that apply only to large well-resourced institutions, without sufcient regard to the requirements of smaller organizations and of individuals. Yet it has become abundantly clear that there is a need to engage all and to approach the issue at all levels. To date, the responses to the challenges of digital preservation have largely been developed by large, well-resourced libraries and archives. So, what can be done about digital preservation in smaller institutions? How can small libraries and archives respond to the challenges of preserving digital materials? The issue here is one of scalability: scaling down, as well as scaling up. Some guides are starting to be produced, for example a guide to contracting out digital preservation services produced by the DPC in response to its members concerns that smaller organizations are not as well-equipped to deal with digital preservation as its larger member organizations (Simpson, 2004a).
A close look at the lessons learned by institutions that have been active in digital preservation to see how they could apply in other contexts repays attention. The National Library of Australias experience provides some guidance and assistance. One of its responses is stages of action . . . Our most pressing demands were to make some decisions about what we should try to preserve and to put those materials in a safe place . . . We have formalized these processes into two broad terms: archiving and long-term preservation. For the NLA, archiving refers to the process of bringing material into an archive; longterm preservation refers to the process of ensuring that archived material remains authentic and accessible (Webb, 2002, p.66). This translates into a staged approach: start now, do what you can now, and then consider the possibilities. Other experiences of the National Library of Australia provide guidance for smaller institutions. Its migration experience is one area where the benets of its expertise have already been provided in the form of guidelines for low-end users, for instance a staff paper Practical Advice for Preserving Publications on Disk, a basic guide for beginning the process of managing and preserving physical format digital publications (Woodyard, 1999). Similarly, the experience encapsulated in Woodyards description of the Librarys migration experiences (Woodyard, 1998) can be applied in smaller institutions. The National Library of Australias experience with CD-R provides ready information for wider dissemination about what types of CD-Rs are best to use, and advice about testing procedures and why they are important. Lessons from the PANDORA experience applicable to smaller institutions are less obvious, but could include guidelines for determining signicance of sites aimed at the local level (for example schools, local history collections, and company web sites) which would assist in raising awareness and in developing selection policies, an essential early step in digital preservation. Stakeholders have expressed interest in learning from the experience of the National Library. One is CAUL (the Council of Australian University Librarians) whose report at the 2002 UNESCO Regional Consultation on the Preservation of Digital Heritage specically requests best practice models . . . as well as guidelines and notes that leadership from the National Library of Australia (NLA) is welcomed (Horn, 2002). The publication of UNESCOs Guidelines for the Preservation of Digital Heritage (UNESCO, 2003), prepared by the National Library of Australia for UNESCO, is an important step towards disseminating the lessons learned from practice. Although these Guidelines are aimed at preservation managers and their audiences are dened as policy makers, high level managers, line managers and technical practitioners (UNESCO, 2003, pp.1819), they include suggestions aimed specically at those people struggling to set up programmes with extremely limited resources (UNESCO, 2003, p.20); many of its chapters have a section headed For preservation programmes with few resources. While they are undoubtedly valuable, it is debatable that they will reach two important groups needing advice on digital preservation small libraries and organizations, and individuals. We can readily envisage a local history collection in a small public library. Such collections are of considerable signicance to the area they serve. We can
Research and digital preservation 193
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
also realistically expect that this library will be offered photographs of its locality in digital form; indeed, if this has not happened already, it will happen in the near future. Similar situations apply to a wider range of institutions, for instance the one-person special library, or the small archive, perhaps staffed largely by volunteers, or the school library, many of which are already attuned to the need to preserve the schools heritage and are nding that much of it now resides on servers or on short-lived media. What these small institutions have in common are digital objects in their collections, staff that are time poor, and very limited resources. These institutions urgently need guidance in the form of precise directions that are concise and able to be readily implemented. The challenge is to codify the lessons learned into guidelines and action lists for smaller institutions. The National Library of Australia has in fact produced such guidelines in their Recommended Practices for Digital Preservation (National Library of Australia, 2004?) and in an earlier guide, First Steps in Preserving Digital Publications (National Library of Australia, 2000?). Guidelines have also been issued by other organizations, for instance recommendations for creators of and those interested in electronic literature (Montfort and Wardrip-Fruin, 2004). There is denitely still room for more. Many digital preservation strategies are centred around digital mass storage systems (noted in Chapter 8). These are major investments requiring large initial outlay and signicant ongoing costs. There is as yet no equivalent for smaller institutions, although the personal digital mass storage system one that is affordable by small libraries and archives has been promoted (Schller, 2000). Until solutions are more universally scalable, what should small libraries and archives who want to begin digital preservation do?
Research and digital preservation

Digital preservation is characterized by a high level of research activity. This is evident in many of the initiatives and collaborations described in Chapter 9. All of them conduct research, or have done so; in some of them, research has a very high priority. Early examples include research commissioned by the British Library to compare methods and costings (Hendley, 1998) and the National Library of Australias Draft Research Agenda for the Preservation of Physical Format Digital Publications (National Library of Australia, 1999a). Some of the most prominent examples are: CEDARS and CAMiLEON, both established as research collaborations; initiatives that were established specically to test strategies and services, such as the Digital Preservation Testbed; and others where research is one major activity along with others, such as the Digital Curation Centre and ERPANET. The high levels of research activity in digital preservation can be explained in part by the novelty of most of the ways of thinking and activities that are needed. Few of the pre-digital paradigm approaches are transferable and (with the exception of the experience of the data archiving community) there was no signicant body of knowledge to fall back on. Another reason is the urgency of the issues: these have, Lavoie suggests, motivated an ambitious, communitywide research agenda aimed at an improved understanding of what digital preservation entails (Lavoie, 2003, p.3). Certainly the engagement of a large number of stakeholders the community-wide support noted by Lavoie is
clear, collaboration being a hallmark of digital preservation, as Chapter 9 notes. Research has resulted in many responses to the technical challenges, such as preservation metadata schemes, increased understanding of the longevity of media, and requisite architectures for digital repositories. Other challenges, too, have been addressed: the social challenges, such as a better understanding of risks, and increased knowledge of authenticity requirements. Statements about the digital preservation research agenda abound general statements, those that are sectorally focused, and some that have a national focus (such as Kerry, 2001 for Australia). One of the general statements is Smiths 2002 identication of ve challenges for digital preservation (policy, benchmarking cost models, awareness building, networking, and technical and research issues), all of which with the possible exception of awareness building require significant research activity. Smith expands the list in considerable detail (B. Smith, 2002, pp.136139). Two others are apparent in the joint US National Science Foundation (NSF) and Library of Congress report, Its About Time: Research Challenges in Digital Archiving and Long-term Preservation (Workshop on Research Challenges in Digital Archiving and Long-term Preservation, 2003), and with the joint NSF and European Unions DELOS report, Invest to Save: Report and Recommendations of the NSF-DELOS Working Group on Digital Archiving and Preservation (NSF-DELOS Working Group on Digital Archiving and Preservation, 2003). These agenda demonstrate a high degree of consensus. An early statement (Bennett, 1997) posed some key questions that are still valid and can usefully motivate our current research: Why? What is the rationale for preservation? When an object is retrieved from the archive will it still be valuable in 50 years time? Will it still be recognizable and comprehensible? . . . what costs are non-discretionary, how do they apply to an items life-cycle in archive, and when will costs start to be discretionary? What benets are measurable, how can they be achieved, and who can be tasked with capturing them? How much? What contextual information is necessary for preservation? . . . what contextual information is sufcient, so that when it is retrieved it can be interpreted correctly? How the object will eventually be accessed, and for what purpose, how will this affect the approach to preservation? . . . How? What are the preservation processes procedural needs in order to achieve a long term archive? Who are the stakeholders, who will inuence the way the archive is built up and managed? What quick, cost-saving routes are there, which do not adversely affect the quality of the archive? What safety nets exist which can provide a fall-back for the archive should accidental loss or deliberate sabotage to the archive occur? Where? While technology is in a state of continuous transition, when will technology be resilient and stable enough for any item to be assured of its long term preservation? (Bennett, 1997, pp.911). Hedstrom in 1998 noted four areas for research that are likely to be fruitful: storage media, migration, conversion, and management tools (Hedstrom, 1998, pp.197200). More recently, the outcomes from the joint US NSF and Library of Congress workshop to consider research challenges in digital preservation provide a comprehensive map of the digital preservation research agenda based on four themes:
Research and digital preservation 195
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
1) Technical architectures for archival repositories: specication, system and tool development, pilot implementation, and evaluation of repository models; develop a spectrum of repository architectures; develop a spectrum of digital archiving services; alternative repository models and interoperability; scalability and cost 2) Attributes of archival collections: articulating and modeling of curatorial processes; developing appropriate preservation methods for diverse digital objects and collections; aggregation of items and objects into collections; decision models 3) Digital archiving tools and technologies: acquisition and ingest; managing the evolution of tools, technology, standards, and metadata schema; naming and authorization; standards and interoperability 4) Organizational, economic, and policy issues: metrics; economic and business models (Workshop on Research Challenges in Digital Archiving and Longterm Preservation, 2003). The joint working group of the NSF and the European Unions DELOS Programme outlines an ambitious agenda. It groups its recommendations for future research into three areas: 1) Preservation strategies: emerging research domains (including research into repository models, software and format repositories, archival media, salvage and rescue, accelerated aging) 2) Re-engineering preservation processes (including research into modeling preservation processes, automation of processes, scalability, distributed and grid storage) 3) Preservation of systems and technology (research into areas such as formats, managing complex and dynamic digital materials, automated metadata creation, acceptable loss) (NSF/DELOS Working Group on Digital Archiving and Preservation, 2003, pp.v-ix). This report noted that three research areas are most likely to have the greatest impact: self-contextualizing objects, metadata and the development of ontologies, and mechanisms for preservation of complex and dynamic objects (NSF/ DELOS Working Group on Digital Archiving and Preservation, 2003, p.ix). There are frequent calls for more reports on practical experiences of digital preservation. The predominant theme of a 2004 DPC forum about the global context of digital preservation was the need to develop and share practical experience (Digital Preservation Coalition, 2004). Lynch suggests that the digital preservation experience to date is most applicable to the largest organizations and so for smaller organizations at the state and local government level, and for corporations of all sizes, we need a research agenda to give guidance and assistance for those who manage these valuable resources; we need to offer them affordable and implementable approaches, and supporting systems. There is a place not just for analytical research but also descriptive research (Lynch, 2004). Although much research has been carried out, there is much more to be done. There are also some notable gaps in research carried out so far. Lavoie notes one of the most signicant: as digital preservation moves beyond the realm of small-scale, experimental projects to become a routine component of a digital assets
life-cycle management, the question of how it can be shaped into an economically sustainable process begins to overshadow other concerns . . . Digital preservation can hardly be classied as a new topic anymore, yet we still nd ourselves not very far from the beginning in terms of exploring its economic ramications. No systematic study of the economics of digital preservation has yet emerged (Lavoie, 2003, pp.4142). Research into digital preservation continues, and will continue into the foreseeable future, to be a pressing requirement if digital preservation is to make real progress.
Conclusion the future of digital preservation

Although there is still considerable wariness and skepticism surrounding our ability to maintain and provide access to digital materials in the long term, it is lessening. One demonstration of this is the 2004 Association of Research Libraries statement Recognizing Digitization as a Preservation Reformatting Method, in which the Association of Research Libraries endorses digitization as an accepted preservation reformatting option for a range of materials. Although this report recognizes that there are still many issues associated with the longterm preservation of digital materials, it acknowledges that the means for short-term preservation are available, that considerable activity is taking place to develop long-term solutions, and it concludes that libraries cannot wait for these solutions to be completely settled before testing the waters. Therefore, we must be prepared for persistent technological change (Arthur et al., 2004). Similar points have been made by others, for example Woodyard who four years earlier noted that at the National Library of Australia we have . . . analyzed the situation, recognized problems and made a good start on implementing some solutions. We have not waited until they are all thoroughly tested and proven because we believe there is limited time to nd solutions due to the rapid rate of technological change (Woodyard, 2000). Few are brave enough to attempt to prophesy what will happen in the future, but one attempt was the outcome of a consensus-gathering exercise carried out in 2004 by DigiCULT, an EU-funded technology watch project. Comments were solicited from a wide range of international experts. One area reported on is persistent and perpetual access (Geser and Pereira, 2004, p.58). The experts suggestions are grouped into four time periods: 2004 (current limitations/ barriers), 20052009, 20102014, and 2015 and beyond. From 2005 to 2009 we can expect maturing base technologies and procedures, including many more Open Source applications; the examples given include software for unattended reformatting, monitoring, and risk management. 2010 to 2014 will see the availability of more powerful distributed preservation services. From 2015 and beyond we might expect digital ontologies that allow all aspects of digital materials to be described in ways that assist their preservation. In the next 10 to 15 years we can expect to see techniques and software developed for digital preservation in smaller organizations (Geser and Pereira, 2004, p.64). Digital preservation has now reached a point of maturity where its requirements can be codied. At the end of a comprehensive overview of archiving in
Conclusion the future of digital preservation 197
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
the digital age, Tibbo provides seven dicta that need to be met to secure the future of digital preservation: 1. Society must recognize, understand, and actively support the efforts to solve the challenges of digital archiving 2. The challenges that electronic records, and more broadly, digital objects, pose in respect to their long-term preservation and access must be fully explored as a priority research and development area that receives both strategic planning and extensive funding 3. Preservation and access must be viewed as having an inseparable relationship 4. Archivists, information scientists, librarians, policy makers, and computer scientists must address the full range of issues integral to digital preservation in a coordinated and collaborative fashion 5. The information technology industry must produce tools to support digital preservation and access 6. The notion of responsible custody of digital assets must pervade society; digital archiving must become ubiquitous 7. Archivists must take a leading role in educating society regarding digital preservation (Tibbo, 2003, pp.4352). Are we, as information professionals, capable of meeting these challenges? Do we have the will, the skills, and the inuence to manage this at the institutional level? Or is the best chance for large-scale, affordable digital archiving likely to be linked to individuals needing to preserve their own digital materials, with the consequence of affordable software solutions being readily available? To this end Microsoft is developing software to offer as part of its Microsoft Ofce suite, in its MyLifeBits project: MyLifeBits is a lifetime store of everything . . . Gordon Bell has captured a lifetimes worth of articles, books, cards, CDs, letters, memos, papers, photos, pictures, presentations, home movies, videotaped lectures, and voice recordings and stored them digitally. He is now paperless, and is beginning to capture phone calls, television, and radio (MyLifeBits Project, 2003). Microsoft is here linking in to the larger trend of life caching (Life Caching, 2004), but, while this is currently in fashion, it may not be a lasting trend. If it does persist, it could be the lever that digital preservation so urgently needs to generate public interest, leading to public support and, eventually, adequate resourcing of those institutions that are charged with the responsibility of maintaining access to information and cultural heritage, regardless of its form, into the future.
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Appendix: Case Studies
These case studies provide examples of Australian digital preservation practice. They are drawn from a range of activities in organizations of varying sizes and with varying resources. Case Study 1 examines digital storage at the National Film and Sound Archive and illustrates the topics of media longevity (Chapter 3) and migration (Chapter 8). PANDORA, the National Library of Australias web archiving initiative, is examined in Case Study 2. This case study illustrates many of the topics found in this book, in particular selection (Chapter 4). Case Study 3, which notes digital preservation activities at the National Archives of Australia, illustrates the topics of normalizing and XML (Chapter 8). VERS (Victorian Electronic Records Strategy) is the subject of Case Study 4; it illustrates the topic of encapsulation (Chapter 8). Case Study 5 describes migration at the Department of Family and Community Services and illustrates the topic of migration (Chapter 8). PADI (Preserving Access to Digital Information) is the topic of Case Study 6. This case study illustrates collaborative activity (Chapter 9) and the importance of education and raising awareness (Chapter 10).
Case Study 1 Digital Storage at the National Film and Sound Archive, Australia
The National Film and Sound Archive (NFSA), previously ScreenSound Australia, is Australias national audiovisual archive. Established in 1984, its role is to collect, store, preserve and make available screen and sound material relevant to Australias culture. This case study is provided by Matthew Davies. As a modern collecting institution, the National Film and Sound Archives involvement in digital storage technology is inevitable. The proliferation of digital formats for recording and distribution has resulted in a large body of information that has as its original form a digitally encoded signal. The most
200 Appendix: Case Studies
powerful tools we can use to restore degraded historical material depend on manipulation of digital data, and all predictions for the future indicate that storage and distribution of sound recordings will ultimately be carried out exclusively in the digital domain. The NFSAs earliest operations with digitally encoded sound depended on technologies which were not developed specically with the requirements of archiving in mind, so choices had to be made about the suitability of available media for the various functions undertaken by the Archive. More recently, a digital archiving system based on Quadriga audio archiving workstations and Storagetek digital mass store systems, integrated with the Archives MAVIS database, has been implemented. What then are the reasons behind the choices that have been made at the NFSA and what affect has the system had on its operations?
The standard
The widespread acceptance of 16-bit quantization and 44.1kHz sampling rate demands that any system used for archiving meets or exceeds those specications. Compressed or otherwise inferior systems would compromise the requirement to maintain the integrity of material in the collection. Systems with higher quantization and sampling frequencies are increasingly being adopted for more critical applications. Initially the NFSAs operations depended on Digital Audiotape (DAT) and CD-R and conformed to the 16 bit 44.1 kHz system, but, with the introduction of the audio digital archive, a range of standards for different materials have been implemented, as shown in the gure below.
Source material Audio Cassette RTR Audio Tape recorded at 7.5 IPS (19 cm/s) RTR Audio Tape recorded at 15 IPS (38 cm/s) Audio disk and cylinder records Digital audio recordings Standards for distribution copies Standards for browsing (reference) copies Quality requirement Low to Moderate Moderate to High High to Very High Very High Variable Moderate to High Low Sampling rate 48 kHz 48 kHz Quantization 16 or 24 bit 24 bit Compression No No
96 kHz
24 bit
No
96 kHz As per the original recording 44.1 kHz 44.1 kHz
24 bit As per the original recording 16 or 24 bit 16 bit
No No No MP3 at 128 kbps
Standards for Different Materials at the National Film and Sound Archive
Case Study 1
201
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Functions and requirements

One of the NFSAs main criteria for accepting material into the collection is that the recording should be the most original or earliest generation possible. With digital recordings, the absence of any generation loss makes this less critical, so if a choice of formats is offered, the most stable, standard format is preferred. If the material has been published on CD from a 16 bit digital master then the stability and wide acceptance of the CD format make it a natural choice for acquisition. The Archive has accepted a variety of formats of master or oneoff recordings, including PCM F-1, 701 and 501, 1610, DAT and CD-R, but to date it has not acquired many recordings on formats such as Sony PCM-900 Masterdisc where the digital encoding exceeds the CD standard. Sound preservation operations require copies to be made of the heritage material in the collection for a variety of reasons. These would have been initially recorded onto DAT tapes prior to introduction of the digital archive but are now captured as les on the audio workstation. As sonic restoration, editing or compilation work proceeds, new les are created for any intermediate copies and the nal recording is saved on the mass store and accessioned into the collection management system. Analogue preservation copies are no longer made as a standard practice, so, for long-term survival and accessibility of historic and contemporary recordings, the NFSA now depends on a strategy of migration of digital data. Factors inuencing the choice of media for this purpose are cost of blank media, data capacity of the media, consequent cost of copying, stability of the media, and the expectations of continued availability of media and hardware. In the absence of an ideal digital preservation format, the Archive previously used DAT and CD-R for digital recording with a view to migrating the data as more suitable systems become available. With the introduction of the digital archive for audio materials the Archive has moved to a system employing LTO Ultrium data tapes, which are regularly copied for backup. Analogue recordings at high risk due to deterioration or obsolescence are given the highest priority for copying. After any necessary conservation work the original recordings are re-played and digitized using high quality analogue to digital converters manufactured by Prism and Digital Audio Denmark. Safety copies are made of unstable originals at the earliest opportunity in the copying process to ensure that any damage to the original in the copying process will not result in the recording being lost. Digital les are created according to the standards shown above, stored on the mass storage system and accessioned into the MAVIS collection management system. Recordings in the Archives collection on obsolete formats like PCM F-1 are also queued for copying, and any recordings received on non-standard digital formats are being added to this queue. DAT tapes of restored recordings made by the Archive have been copied to CD-R for access purposes, but also as a safeguard against the failure of the DAT media. The Archive began with the oldest DATs in the collection, which were then approaching six years, and, at this stage, only a very small number of irrecoverable errors have been noted. A checking program has been instituted for DAT tapes, and tapes will be copied when their error rates approach a level of risk. Access copies on CD-R are generated as part of sound preservation programmes, and targeted compilations are also produced to meet existing and
expected demand. CDs provide a suitable level of audio quality for these valueadded materials, and the convenience that has ensured their popular success is equally relevant to their use in client servicing. CDs are sufciently familiar so as not to present difculties to clients without a technical background, and are sufciently simple and robust to minimize the likelihood of damage when used by people without operational skills. While not providing truly random access, the CD, with indexed tracks, provides a convenient package for browsing, and this also simplies copying. Clients of the Archive can request copies of recordings on the media of their own choice. The CD is increasingly popular with the Archives private clients, and for some professional uses. DAT has been quickly superseded as a distribution format for production uses, and production quality audio is now generally supplied on CD-R or via le transfer. MP3 les are produced for any content with an anticipated high level of access demand. These les are stored on the mass storage system and accessible directly from the MAVIS record. Where copyright has been cleared these les are also accessible from the National Film and Sound Archives online catalogue on the web. In all cases the Archive retains the original analogue recordings due to their value as primary source materials.
The media
The National Film and Sound Archive currently holds approximately 15 000 massproduced CDs. In general, the mass-produced CD performs well as a medium for acquisition, preservation and access to audio recordings. The stability of properly manufactured disks would seem to ensure their survival beyond the obsolescence of the format, and the relative number of disks with manufacturing faults is comparable to analogue disks. Cheap replay equipment is available, and the reproduction process poses minimal danger to the disks. CDs are widely accepted, easy to use, have simple storage requirements and are really cheap to make. In 1994 the National Film and Sound Archive standardized on recordable CDs as an access medium for material in high demand, and production of CD-Rs continues, although the NFSA is moving to use MP3 les over its network as its preferred access method. The Archive has about 1500 CD-Rs in the collection, and there is an expectation that this will increase by about 100 next year. Reductions in the price of blank CD-Rs have increased the cost effectiveness of CD-R so that it compares favourably with other digital storage solutions. Compilation of CDs has proven to be labour-intensive, but becomes easier when one has a more complete digital signal chain. Wide acceptance of the format has ensured that the Archives efforts in producing disks have been well received, and this is also an important factor for the long term. DAT has been used by the Archive for the past 10 years as an intermediate or working copy, and the Archive now holds about 3400 DAT tapes. Most of the DATs produced in-house are recordings that have been sonically enhanced by a variety of systems. These DATs were used as a source for client copies and compilations, but are now largely redundant for this role. A small subset of the DAT collection comprises original unique recordings, mostly eld recordings, and these will be copied to the digital archive within a few years. Total holdings of other digital formats are less than one per cent of the Archives collection. These recordings include location and oral history recordings and unpublished master tapes. The proliferation of semi-professional and
Case Study 2
203
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
special purpose digital formats poses a signicant challenge to collecting institutions, as the range of reproduction equipment required steadily increases. Many of the recordings on these media are unique, and they are likely to remain in the hands of private individuals until the effects of age begin to cause problems with the recordings. By the time any signicant items are acquired by a collecting institution, it is possible that reproduction equipment may no longer be available, and the complexity of digital reproducing machines makes it unlikely that the Archive will be able to easily re-create players for formats which have failed in the marketplace.
Strategies
The National Film and Sound Archives effort began with transferring recordings onto the two most widely accepted physical digital formats DAT and CD-R and has now progressed to a centralized mass store-based system. Digital recordings generated in the course of earlier preservation and access programmes serve to replace recordings on obsolete formats or unstable carriers, and will be used as a source for digital migration to the digital archive in the near future. The Archives growing digital holdings will demand regular sampling, monitoring, and error checking to plan and implement migration or copying programmes. Cooperation between institutions and an active involvement in setting and maintaining standards will be important to the Archives success. To this end the NFSA is involved in cooperative projects with the National Library of Australia, the International Association of Sound and Audiovisual Archives (IASA), and the Association of Moving Image Archivists (AMIA) to establish standards and guidelines for digitization and management of audiovisual heritage. The publication of IASA TC-04 Guidelines on the Production and Preservation of Digital Audio Objects in August 2004 (International Association of Sound and Audiovisual Archives Technical Committee, 2004) represents an important milestone in this cooperation.
Case Study 2 PANDORA (Preserving and Accessing Networked Documentary Resources of Australia)
The National Library of Australias PANDORA (Preserving and Accessing Networked Documentary Resources of Australia) Project is examined in this case study. It illustrates many of the topics covered in this book, in particular selection (Chapter 4). Approaches to preserving web sites fall into two camps. One, exemplied by the National Library of Australias PANDORA, is based on selection of web resources in order to preserve and provide access to them in the long term; the other, the whole domain harvesting approach, is exemplied by The Internet Archive. The PANDORA archive is the only real example of collaboration and coordination in the collection of digital information resources that Beagrie encountered in his four-country survey of best-practice digital preservation (Beagrie, 2003). While it is not the only example of web archiving the Internet Archive and the National Library of Canadas Electronic Publications Pilot
Project are other pioneering projects it is one of the earliest and has provided experience on which other institutions have drawn. PANDORA, sometimes also known as Australias Web Archive, began life as a project to collect and preserve the digital national heritage in the form of online publications. It is based on the conviction that the Library has a responsibility to collect and preserve the published national heritage, regardless of format (Webb, 2002, p.69). From a small start in 1995 with 36 titles, PANDORA is now an archive of selected, signicant Australian web sites and web-based online publications, an operational National Collection of Australian Online Publications (Webb, 2002, p.70). It is based on the principles of retaining the look and feel of the material it archives, of being selective rather than comprehensive, of selecting titles which are available in online format only, and of creating records of archived material for the National Bibliography (Webb, 2000c, pp.184187). As the Library insists that what it collects and preserves must be accessible, this raises resource issues to provide accessibility, so selectivity is considered to be the only realistic approach. Smith (1997) describes PANDORAs early history. It commenced with a modest pilot project in 1995, at the time of the establishment of the National Librarys Electronic Unit, which experimented to see what would work in capturing selected publications. A selection policy for online materials was developed based on the Librarys existing collection development policy: collecting would not be comprehensive, as was attempted for Australian printed materials, but selected items should have signicant Australian content or authorship, and the work should be authoritative with reasonable research value. Permission to archive would be sought from owners. By 1998 procedures were well established and the involvement of other libraries was sought. The State Library of Victoria became a partner in 1998. Although by 1999 the procedures were well established, internet sites had been archived, and access provided to them, there were still challenges: methodological (such as applying selection criteria), resourcing (such as maintaining staff and equipment at satisfactory levels), and technical (the harvesting software and archive management process were giving concern) (Morrison, 1999). A case study of adding the web site Lighthouses of Australia illustrates how cumbersome and time-consuming the process could be (Morrison, 1999, pp.282283). The concept on which PANDORA is based is one of a national distributed archive with other stakeholders playing a role. After the State Library of Victoria had become the rst partner in 1998, other state libraries followed: the State Library of South Australia in 2000, Western Australia in 2001, and New South Wales (which had begun its own web archiving operation earlier) and the Northern Territory Library and Information Service in 2002. PANDORA closely cooperates with the State Library of Tasmania, which maintains its own web archiving operation, Our Digital Island (odi.statelibrary.tas.gov.au). Other organizations also joined: the National Film and Sound Archive (then ScreenSound Australia) in 2000, and in 2004 the Australian War Memorial and the Australian Institute of Aboriginal and Torres Strait Islander Studies. Statistics from June 2004 indicate that the partners are contributing signicantly to PANDORA: the National Library was responsible for about 57 per cent of archived titles, for instance. Collaboration extends to cooperation with publishers and resource creators. The National Library and the Australian Publishers Association have developed a Code of Practice for Providing Long-Term Access to Australian Online
Case Study 2
205
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Publications, and in 2000 the Library issued Safeguarding Australias Web Resources, a set of guidelines providing advice on creating, describing, naming, and managing web resources. Even with its selective approach PANDORA has grown rapidly. In 2000 it had archived over 900 titles (40 gigabytes) and selected about 35 new titles each month. At the end of May 2003 it contained 3936 titles (over 440 gigabytes) and was growing by 141 titles (six gigabytes) per month. About one-third of these titles had been gathered more than once to illustrate the changing nature of web sites, and many are gathered on an ongoing basis to capture new issues of serial publications. PANDORA had already clearly proved its worth: many of the sites captured by that date, such as the Sydney 2000 Olympic Games web site, are now only accessible through PANDORA. In August 2004 the number of titles had grown to 6608 (702.1 gigabytes) and by October 2004 to 7000 titles, of which 150 had access restrictions placed on them by publishers, and an estimated 7.4 per cent of which were no longer available on the live web. At the end of February 2005 there were 7813 titles archived (802 gigabytes). As noted above, the big debate in the web archiving eld is whether to take a selective approach or a whole domain harvesting approach. PANDORA believes in a selective approach to archiving, choosing material that most strongly represents Australian social, intellectual and cultural life. Selection is based on well-dened selection guidelines that t within the Librarys overall collection development policy. The National Librarys policy of selective acquisition is inuenced by resource issues and by its goal that what it collects and preserves must be accessible. It maintains that this approach, compared with the harvesting approach, allows an emphasis on quality for example, checking each title to ensure it has been fully captured. The selective approach was developed in the early years of PANDORA and was justied in this way: The main reason was the insubstantial nature of most web publishing: the library does not attempt to collect printed ephemera comprehensively, and could see no justication for collecting its electronic equivalents . . . The key point about selectivity: it is our [that is, libraries] normal practice (Morrison, 1999, p.275). There were also issues associated with the Librarys legal ability to collect all resources, so that publishers permissions were required as Australia has no legal deposit legislation for digital publications (Webb, 2002, p.70). The selective approach also allows the Library to devote some resources to quality assurance procedures, checking each title to make sure that it has been copied in its entirety and is fully functional. The selection process used by PANDORA places content before format. To be selected the materials should be about Australia, be on a subject of signicance and relevance to Australia and be written by an Australian author, or be written by an authoritative Australian author. The location of the server from which the material is harvested is irrelevant: it may be in Australia, but need not be. If material exists in both a print version and a digital version, the print version is collected. There is no attempt to collect all versions or editions of sites; rather, PANDORA sets a schedule for harvesting that site on a regular basis (Phillips, 2004, p.3). The placing of content above format causes some issues on occasion, for example there are technical issues about making sure that some
dynamic sites are accessible. About 40 per cent of the sites selected need further work. This is an unusual aspect of PANDORA compared with other web archiving programmes. The selection guidelines are available on the PANDORA web site (pandora.nla.gov.au/selectionguidelines.html). They were reviewed and revised in 2003 and currently indicate six priority categories: Commonwealth and Australian Capital Territory government publications, publications of tertiary education institutions, conference proceedings, e-journals, items referred by indexing and abstracting agencies, and sites in nominated subject areas and sites documenting key issues of current social or political interest (for example election sites, the Sydney 2000 Olympics, and the 2003 Canberra bushres). PANDORAs selectivity does raise questions about the usefulness of this approach. Libraries always face the problem of guessing the future when they develop selection policies, of course, but they must frequently revisit such policies if evidence suggests that changes are required. To take one example, there are signicant discrepancies between the number of instances of the Ansett Australia web site available on PANDORA, compared with the Internet Archive through its Wayback Machine, as noted in the table below. (Ansett Airlines sudden collapse in 2001 caused much anxiety in Australia.) Of course, these gures say nothing about the quality of the archived sites in terms of how functional they are, how many images are captured, and other characteristics.
1996 PANDORA Internet Archive 3
1997 3
1998 4
1999 9
2000 11
2001 4 0
2002 4 11
2003 1 15
2004 2
Archived instances of the Ansett Australia web site at 22 March 2005
Smith (2005) has investigated a sample of sites in more depth and makes the observation that of her sample, selected as having some Australian signicance, after nearly ten years a large proportion . . . can still be found online, even though they are not at the original URL. However, accessibility continues to drop, and, in another ve years or so, only 50% of the titles are likely to be found online. The Internet Archive provides the only mechanism for ongoing access to the content of the majority of these titles. What would be of interest are some studies, from the users perspective, comparing PANDORA with the Internet Archive for the presence of Australian material. This study would, of course, need to recognize that the scope of what each collects is different, as are the resources available to each organization, and the quality of what each captures. The National Library is of course well aware that there are disadvantages to PANDORAs selective approach. Webb noted in 2002 that while the Library recognizes many advantages in taking a more comprehensive approach to Web archiving, we have yet to be convinced that such advantages outweigh the benets of quality control and accessibility that we have been able to achieve only
Case Study 2
207
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
while collecting selectively. He indicated that the National Library would like their selective approach to be complemented by periodic capture of more comprehensive snapshots of the Australian domain and in 2001 had employed a consultant to investigate this possibility (Webb, 2002, p.70). Phillips summed up both the disadvantages and the advantages. The advantages are that the archived materials are quality assessed and are functional, they are fully catalogued in Australias National Bibliographic Database and are therefore accessible, items can be made available immediately because permission from publishers has been negotiated, and publicly inaccessible sites are gathered and made accessible. The disadvantages noted include that selection involves making a subjective judgment about value and future usefulness, that resources are taken out of context and links to material referred to are not functional, a tiny amount (estimated at less than one per cent) is collected, and the process is labour-intensive and has a high unit cost (Phillips, 2004, p.1). In 2005 the National Library of Australia will consider strategies to increase its rate of harvesting web resources into PANDORA. The possibilities include thematic or domain harvesting, developing different models in which partners can contribute, and active deposit of digital publications by Australian publishers, where the initial focus will be on Commonwealth government online publications. Organizationally, PANDORA is integrated into the selection, acquisition and collection management processes of the Library. It is managed by the Electronic Unit, which forms part of the Collections Management Division of the Library. Catalogue records for every web site are included in the Librarys catalogue, the National Bibliographic Database, and on the PANDORA web site, providing multiple paths for resource discovery. PANDORAs system architecture and workows are described by Koerbin (2004). Initially the web harvesting software WebZIP was used, later replaced by HTTrack (Koerbin, 2004). At the beginning an Access database combined with a paper-based system were used for archive management. This was replaced in June 2001 by PANDAS, built in-house and developed after unsuccessful attempts to locate an off-the-shelf system to provide an integrated web-based archiving management system. A new version of the PANDORA digital archiving system was released in August 2002. It included an improved search interface, automated many of the processes, and was web accessible by partners who no longer needed locally mounted software. A successful trial migration of les was carried out to investigate whether there were any issues caused by superseded formatting tags in HTML (Webb, 2002, p.71). Future developments are likely to include development of PANDAS to accept metadata from Kinetica (the National Library of Australias host for the National Bibliographic Database) which will trigger automatic harvesting, and applying the International Internet Preservation Consortium (IIPC) tools for harvesting, when they are available. PANDORA has also contributed signicantly to the development of digital preservation awareness and knowledge in Australia and actively collaborates with other web archiving initiatives, most recently through its participation in the IIPC (noted in Chapter 9). PANDORA has also provided an excellent training ground for the National Library to work out how to select Australian online publications and how to manage and provide access to them over time. Through it the National Library has been able to achieve tangible web archiving outcomes while continuing to develop, extend and rene its systems and workows (Koerbin, 2004, p.23). Phillips notes that
one of the most important lessons we have learned from this Project is Take Small Steps. Initially it was very hard to know how to break the chaos barrier. Where could we begin? By taking one rst small step developing selection guidelines for what we would want to collect and preserve, if only we could nd the wherewithal to do so we started on a long steep learning curve (Phillips, 2000). PANDORA is heavily used, according to survey results noted by Phillips (2004, pp.67), who adds that there is a little bit of anecdotal evidence to suggest that PANDORA is used not just for unique material, but because it is recognized as a reliable and well-organized source for high quality Australian content. This might be wishful thinking. PANDORAs focus is still narrow, and its revised 2003 selection guidelines, suggest one commentator, restrict its collecting even further (Smith, 2005). It specically excludes more innovative forms of publishing such as blogs and other forms that the web is giving rise to which have no equivalent in the print world (Smith, 2005). However, PANDORA is likely to widen its focus. A paper summarizing the Archiving Web Resources Conference held at the National Library of Australia in November 2004 observes that the three models of collecting web publications are moving towards similar models. (The three models are the global model, represented by the Internet Archive; the regional model, exemplied by the Swedish Royal Librarys Kulturarw3 project; and the selective and thematic model, exemplied by the National Library of Australia, the British Library, and the Library of Congress.) The paper notes that what appears to be emerging is a dual strategy of occasional complete snapshots and more selective, in depth collection of individual sites of interest (Archiving Web Resources, 2004, pp.2627).
Case Study 3 Changing Preservation Activities at the National Archives of Australia

The National Archives of Australia (formerly Australian Archives) has for many years been concerned with the preservation of records in electronic format and, in common with other national archives, was an early player in digital preservation. This case study describes the way in which its digital preservation activities have changed over the last three decades. It is based in part on a conference paper presented by Stephen Ellis (Ellis, 2003). It illustrates the topics of normalizing and XML (Chapter 8). In common with other national archives, the National Archives of Australia was confronted at an early date with the preservation challenges posed by digital materials (or electronic records, the term more commonly used in the recordkeeping context). An early experience was with the electronic records resulting from petroleum exploration. Australian law requires that a copy of all exploration results, including seismic data on tape, is provided to the Commonwealth government. Over 600 000 tapes of seismic data in a wide range of formats came under the control of the National Archives of Australia in 1974, posing issues such as physical decay of the tapes, and the inability to appraise them because equipment was not available to access them. Lessons were learned from this experience, such as the need to be able to access the data in order to effectively
Case Study 3
209
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
appraise and manage them and assure their authenticity, and that all stakeholders must be clear about their roles and responsibilities, which need to be formally agreed to at the outset (Yorke, 1997, p.70). This material has since been transferred to Geoscience Australia, which has the skills and knowledge to appraise, maintain and use it. The Archives provided advice to government agencies, such as When Its Gone Its Gone!!: Keeping and Disposing of Information on Ofce Automation Systems and Personal Computers (1998) which advocates care in deleting les, sets out criteria for deciding what to retain, and recommends printing nal versions of documents and emails. Other advice was produced as Managing Electronic Records: A Shared Responsibility (1995) and the more detailed Keeping Electronic Records: Policy for Electronic Record Keeping in the Commonwealth Government (1995) which articulated the strategy of keeping electronic records in agency custody. In the early 1990s the National Archives of Australia declared publicly that the problems presented by electronic records were not soluble using contemporary methods. The solution they proposed for government was to leave such records with their originators in what was then a controversial tactic of distributed custody (Ellis and Stuckey, 1994), sparking years of intense debate between post-custodialists and their opponents (described in Boadle, 2004). In 2000 the Archives again changed its custody policy and resumed accepting transfers of electronic records. Some history of what the archival world was like when this series of vehement debates erupted, and of developments in archival theory and practice that led to the present solutions, can assist us to understand these changes. The rst half of the 1990s was a period of intense interest in archival circles about the problems presented by electronic records, and even about what constituted a record. The nature of the problem had by that time become clear to even the most technologically unsophisticated archivist electronic records were inherently ephemeral because of the short time for which the software that created them was available in forms that were useable, an evanescence that was almost equally shared with the hardware upon which such software had to be operated and the media on which they were written. But at the beginning of the decade, even the terms hardware and software were not in the common vocabulary of archivists. Several quite long-established organizations had custody of large bodies of various forms of electronic data and even preservation of the physical media was beyond their resources. Further difculties were added by conceptions of what constituted a record. At the University of Pittsburgh the Americans, ignoring centuries of European archival science and their own history of ofce management, embarked on an ambitious project to reinvent the wheel from scratch and establish the Functional Requirements for Authentic and Reliable Records Project. Provoked by the callow tone of this work, protagonists like Luciana Duranti and her team at the University of British Columbia entered the lists to remind both the Americans and the Australians of the centuries-old fundamentals of archival science and their relevance to the preservation of electronic records. Distinguished Canadian archivists such as Terry Eastwood and Terry Cook contributed to the theoretical debates, while long-serving members of the profession such as NARAs Charles Dollar continued to bring a practised perspective to the forum. This period can be seen as one of considerable development, many of the issues that were debated having passed into common practice. From these
debates came the fundamental practical work that led directly to the Australian Standard for Records Management, which in turn led to the International Standard. In Australia this development owed more to the practical requirements provoked by the problem of addressing electronic records than it did to the theoretical work at Pittsburgh and the University of British Columbia. Similarly, within Australia the conceptual work that led to the explicit articulation of the continuum metaphor to replace the life-cycle metaphor for the management of records and archives came directly out of the debates about the way forward with electronic records. Australian archivists recognized clearly from their experience that the life-cycle approach offered no solution to the electronic records problem. By the time electronic materials had reached the secondary stage stage of the life-cycle model they had moved beyond the point of no return for preservation sustaining both software and hardware over the time-spans common for such a model was impractical. So Australians turned to the more predictive model originally articulated in Canadian Treasury Board procedures of the 1950s and 1960s for traditional-format records get in early, establish comprehensive recordkeeping systems based on agency business functions, and integrate disposal decisions within the structure of the systems. One practical outcome of this was the DIRKS methodology developed by the National Archives in collaboration with the State Records Authority of New South Wales. These debates and discussions also led to being much more specic about what we perceived as adequate metadata for records, work that led directly to the AGLS (Australian Government Locator Service) metadata standard and its successors. It is likely that much of this work in Australia would not have been done without the stimulus of the electronic records problem and the debates it provoked. Some investigations were productive, and some were less so. Productive investigations included the International Standard, which dened the nature of a record and the essential requirements of systems adequate to capture and keep them. The metadata standards similarly establish fundamental characteristics required for authentic and reliable records. In the DIRKS methodology we have a systematic, accountable and replicable process for implementing this suite of standards. A less productive trail evident in the discussions around the earlier part of the decade revolved around the relationship of records to data and to the applications that created them. The Archives most recent advances in solving the long-term preservation problem have come from adopting a data-centric approach to records rather than an application-based one, an approach that recognizes that records are data and should be dealt with as data. The Archives also believes that another false lead has been the emulation trail the attempt to ensure preservation of records by emulating the software and/or the hardware on which they were created. With the exception of a few peculiar instances, this work has simply transformed the problem of preserving records into one of maintaining emulation software in a constantly changing environment, without in any way addressing the media-dependence issues. The National Archives of Australia believes that, while it does not yet have a once-and-for-all solution, it does have a solution that will serve adequately for at least the next generation and will ensure that authentic and reliable records can be sustained in an accessible state over time within an archival institution. This solution is economically and technologically sustainable and is currently
Case Study 3
211
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
in various stages of being implemented in at least three government archives in Australia. It is the XML-based treatment of records as data objects derived from DIRKS-compliant recordkeeping systems, transferred to the Archives with associated recordkeeping and series-system metadata and maintained by the Archive in a quarantined fashion so as to preserve their authenticity. Descriptions of the standards, process models and procedures for implementing such an approach can be found on the web sites of the National Archives and of the Public Record Ofce of Victoria. Both National Archives and the PROV are currently actively implementing such an approach. The National Archives has developed freely available software, Xena, which enables the normalization of digital records created in most common ofce formats to a platform-independent state that ensure they will be retrievable into the foreseeable future. The State Records Authority of NSW is also in the course of adopting this approach, although they are at an earlier stage. These three archival institutions have agreed to collaborate together to form the Digital Archives and Records Coalition of Australia (DARCA) to harmonize and further develop standards and guidelines for an Australian digital preservation method. The Agency to Researcher Digital Preservation Project brings together the National Archives of Australias current digital preservation activities. The AtoR Project began in 2001 and aims to create a way of accepting Commonwealth agencies electronic records those that are addressed as being national archives and preserving them so that they can be made available to researchers in the future. It has done this by developing a digital preservation strategy and building the digital preservation tools that are required to make this a functioning program. The basic conceptual approach is to get electronic records into appropriate open (ie non-proprietary) and well-documented data formats based on XML, and providing tools that handle these XML formats (National Archives of Australia, 2004). A National Archives Green Paper describes this approach in detail (Heslop, Davis and Wilson, 2002, p.6). It articulates the performance model: In the case of digital records, archivists are not interested in the original records but in capturing and recreating the eeting and temporary performance of that record on the screen when it was viewed (Heslop, Davis and Wilson, 2002, p.8). This report notes some lessons learned from a study of migration and emulation, such as that most of the investment in digital preservation is required at the start rather than in continual emulator maintenance or data conversion. It developed the concept of a records essence as a way of providing a formal mechanism for determining the characteristics that must be preserved for the record to retain its meaning over time (Heslop, Davis and Wilson, 2002, p.13). Five principles underpin the AtoR Project. It 1. must be able to preserve any digital record that is brought into National Archives custody regardless of the application or system it is from or data format it is stored in 2. must determine and preserve the essence of the digital records . . . and recreate their essential performance over time 3. will be based on non-proprietary technologies 4. will minimise the number of preservation treatments applied to each digital record 5. will not limit the accessibility choices of the National Archives or of future researchers (Heslop, Davis and Wilson, 2002, p.1417).
The National Archives of Australias approach is based on the use of XML. It has developed a range of XML data formats (or schema) which it will use as archival data formats. Agency records rst undergo normalization (conversion of the original digital materials into XML), then transformation (the XML is converted into an archival format). Specications of the XML data formats/ schema developed are available on the National Archives web site. The process is handled by the Xena (XML Electronic Normalising of Archives) tool, developed by the National Archives. Version 1 of Xena was released in 2004. It converts a range of le formats to XML representations for longer term access, currently handling MS Word, Excel and Powerpoint, OpenOfce formats, Rich Text Format, several email formats, comma separated les, relational databases, JPEG, GIF, TIFF, PNG and BMP, HTML and web sites, and plain text. The National Archives believes that these formats will cover a high proportion of the les it receives from agencies, and will add more in future. Xenas development as open source software and its availability free on the web (xena. sourceforge.net) assist in making the tool itself archival. Because it is open source and freely available it can be adopted by others and adapted to meet their own requirements. It is implemented in Java, with comments included within the code (Rushbridge, 2004b, p.32), which means that documentation about Xena is integral to it; availability of documentation is essential for ensuring the longevity of software. Rushbridge assesses Xena in this way: among the steps being taken to preserve our cultural heritage, Xena and its development principles form an addition that should not be overlooked (Rushbridge, 2004b, p.33). The National Archives of Australia also makes available policies and guidelines that assist agencies in digital preservation. In 2001 it made available a policy for government departments for keeping web resources (National Archives of Australia, 2001a) and accompanying guidelines, which include technical detail for capturing web sites (National Archives of Australia, 2001b). In 2004 it produced guidelines for digital recordkeeping: Digital Recordkeeping: Guidelines for Creating, Managing and Preserving Digital Records, and a Digital Recordkeeping SelfAssessment Checklist. These stress the need for early intervention for preservation of digital records.
Case Study 4 VERS (Victorian Electronic Records Strategy)

VERS is an initiative of the Public Record Ofce of Victoria, in Australia. This case study, provided by Rachel Salmond, illustrates the topic of encapsulation (Chapter 8). VERS is a strategy not a system of any sort, nor a piece of software. It is a strategy in the sense used in Chapter 6, that is, a long-term plan to achieve a long-term aim. As such it encompasses the development of practices, the identication and implementation of standards and the assembling of appropriate technologies. VERS describes itself, in the brochure that is in itself a part of the strategy (making sure that the strategy is known and understood by those who will be required to implement its outcomes), as a world-leading solution to the problem of capturing, managing and preserving electronic records. VERS is a framework of standards, guidance, training, consultancy and implementation
Case Study 4
213
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
projects, which is centred around the goal of friendly and authentically archiving electronic records (Public Record Ofce of Victoria, 2004). Those who will be required to implement VERS outcomes are those responsible for managing recordkeeping throughout the government of the Australian state of Victoria. Those responsible for developing and promoting the strategy are based in the VERS Centre of Excellence at the Public Record Ofce of Victoria (PROV) in Melbourne. The overriding motivator for the strategy is the Victorian Governments commitment to ensuring that its electronic records are not lost, but are managed in the most effective and cost-efcient way for present and future purposes. The story of VERS reveals an exemplary model of the collaborative nature of a large-scale preservation initiative. The Victorian government, industry partners and academic researchers have collaborated in developing this strategy since its inception in 1996. Ernst and Young Consulting, in conjunction with CSIRO (the Australian governments research body) and the University of Melbourne, formed the Project Team engaged by the PROV to produce the conceptual business case to solve the problem of ensuring the retention of digital records. Their recommendation was based on the examination of techniques which support the separation of content from application. As a result, the team believes it is possible to develop a solution that allows long term access to content and context for the majority of records without the need to migrate the application in which the records was created . . . Accordingly, the solution proposes that records be frozen into a static view/print only format at the time of their creation. The frozen image of the document is captured into a to-be-developed or determined Representation Format through which the same record will be retrieved in years to come with satisfactory and evidentiary quality replicability (Public Record Ofce of Victoria, 1996, p.6 (underlining as in the original)). The project that arose out of this recommendation was charged with proving that a frozen image could be captured. It also involved CSIRO and Ernst and Young services and personnel who demonstrated that capture and archiving of electronic was possible, recommending: that PROV specify a preferred long term electronic record format, based on the format described in this report that PROV specify a minimum metadata set that should be associated with every electronic record, informed by the metadata set used in this project that the Victorian Government proceed to implement an electronic archiving strategy based on a long term electronic record format (Public Record Ofce of Victoria, 1998, pp.34). Encapsulation was chosen as the technique for storing documents in their longterm format, together with the information, in XML, about their context (the when, where, how, by whom of the creation and use of the documents). The encapsulated bundle has a digital signature attached to it to ensure the integrity and authenticity of the document or documents it contains. In April 2000 version 1 of the document (PROS 99/007), containing the standard, specications, and
advices for the mandatory implementation of VERS throughout the state government, was launched. Version 2 was released in 2003 (Public Record Ofce of Victoria, 2003). Specication 1 (System Requirements for Preserving Electronic Record) gives all the functions that a recordkeeping system must support and describes specic requirements: record authenticity and integrity; document conversion; metadata capture; modifying information associated with records and folders; documenting folder and record history; reliability; media refreshing; record export. Specication 2 (VERS Metadata Scheme) denes all the types of metadata required and the elements of each. The self-describing digital object for preservation, called a VEO (VERS encapsulated object) is specied in Specication 3 (VERS Standard Electronic Record Format), which includes details of XML and digital signature requirements. The long-term formats for electronic documents within the VEOs are specied in Specication 4 (VERS Long Term Preservation Formats) as PDF or TIFF; these are seen as providing the best chance of ensuring that documents are in their original form into the future. Specication 5 (Export of Electronic Records to PROV) gives details of the mechanisms for transfer of records for permanent storage at PROV physical mechanism of transfer, the encoding of the VEOs on the physical transfer media, and the mechanism for the indication to an agency of PROVs acceptance of responsibility for preserving upon transfer. The pilot implementation of the strategy began in the Department of Infrastructure in 1999. By 2002 that Department was capturing electronic records into an electronic repository and making available the benets of its experience and the software it had developed to other government agencies as they proceeded with implementation of VERS. Once the capture practices for electronic records had been established and tested, and the roll-out to all government agencies commenced, the next major project in the VERS programme was the development of a repository for longterm storage of those records identied for long-term preservation as archives; processes for the transfer and management of, and for access to these archives also needed to be developed. Tenders were called for the building of a Digital Archive for the PROV, and DMR Consulting, with their partners, Fujitsu, began work on the project in 2004. It is expected that the Digital Archive repository, developed on a Documentum platform that interfaces with PROVs Archives One system, will be ready to receive transfers of electronic records from July 2005. A new web site will also be available as a single point of access for searching and ordering paper and digital archives that are stored by the PROV. Viewing of digital and digitized archives will also be enabled through the web site. Funding of VERS projects has come from the Victorian state government, stemming from $20 million announced during the state election campaign in 1996 for the establishment of an electronic archive system for the state. VERS Program Director, Howard Quenault, has reported regularly on the VERS strategy and has shared the lessons learnt during the implementation of its component projects (for example, Quenault, 2001, 2003, 2004). Several of the lessons relate to the apparent lack of awareness in the administrators of public business of the issues relating to the management and retention of electronic records. These people wanted to get on with the business of doing business and expected that their agencies IT infrastructure would capture the records automatically (Quenault, 2003, p. 4). To raise and maintain awareness of the
Case Study 5
215
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
purpose, process and progress of VERS, the project team has paid particular attention to communicating information about VERS to those who need to know about it in meetings, presentations, training programmes, and industry briefings. This communication has not only been directed at the government recordkeeping community, but also at the IT vendor community to ensure that the electronic document management products they supply to Victorian government departments are VERS-compliant. They have sustained this information dissemination role in the programmes web site (www.prov.gov.au/vers), where all the documents needed for understanding and implementing the strategy can be located.
Case Study 5 Migration at the Department of Family and Community Services, Australia
This case study is based on a conference paper presented by Liz Reuben (Reuben, 2004). It illustrates the labour- and time-intensive nature of even smallscale migration projects, and the benets of automating as much of the process as possible consistent with quality assurance. It also illustrates the topic of migration (Chapter 8). In 2002 it became apparent to the Library and Information Services of the Australian Government Department of Family and Community Services (FaCS), that unless action was taken an important corporate record would be lost. FaCS serves several other federal government departments and its potential client group numbers 35 000, all except 2000 of whom are in several hundred locations across Australia. It provides as many services as possible direct to peoples desktops via their PCs. Much of the work of FaCS and other government departments is governed by three related acts, the Social Security (Administration) Act 1999, the Social Security Act 1991, the Social Security (International Agreements) Act 1999, and associated subordinate legislation. These were based on the Social Security Act 1991 and, earlier still, the Social Security Act 1947. One of the primary tools used to interpret the social security law, explain eligibility for payments, and uphold appeals is the Guide to the Social Security Law, an electronic serial available initially via the desktop or on oppy disk/CD-ROM, and, for the last ve years, via either departmental intranets or the internet. Its historical releases provide evidentiary value for the appeals process and for understanding how policy has developed. Superseded releases were originally kept on oppy disk. Each release of the Guide took up 10 to 14 oppy disks and needed to be installed on a standalone PC in the library to run, partly because of the proprietary software, but also because of the security requirements for staff desktop PCs, which required administrator access to install the releases. Releases of the Guide were produced using a proprietary software package, Epublish, dating from the early 1990s; it was written in Visual Basic and consisted of two tools: a desktop publishing program, and a browser. Over its life it became an extremely sophisticated publishing product that enabled hypertext linking within the Guide and between the Guide and its support documentation. At around the time of the major rewrite
of the Guide in 1998/99, Epublish moved to publishing .doc les directly in HTML, including making the links within and between documents, and to documents held on other internet sites. The company producing and supporting Epublish decided that the product was outside their core area of expertise and that they would no longer support the product from March 2003, although this date was later extended. The demise of Epublish was a catalyst for action. In 2002, funding was obtained for a project to extract the data from the proprietary software format of all the releases held, and then migrate it to formats that would support archival best practice and provide greater accessibility. Releases from 1994 were available, except for several gaps, of which the most signicant was 18 months in 1996/7 for which a box of oppy disks had been lost during a move before the material was lodged with the Library; a few other releases had either been lost or had become corrupted. The project team of one began by looking for missing releases, researching best archival practice for digital media, ascertaining the relevant records management and archival practices of the organization, and planning for client input into the nal product. In late November 2002 it was discovered that the owners of Epublish had Rich Text Format (RTF) copies of a product that incorporated Guide releases. This product had been released quarterly, rather than fortnightly. This discovery allowed some of the largest gaps to be lled. Also located were duplicate copies of other releases and some releases from 1991 to 1993. Investigations in December 2002 led to the National Archives of Australia (NAA), which provided several important directions to explore. The three formats preferred by NAA for digital archiving were XML, HTML and PDF, each of which is a format with open specications. NAA also suggested that attention be paid to the process of extracting the data, and to establishing the essence of the document what was vital to be retained in the migration process, and what could be altered or lost without affecting the look and feel or the evidentiary value of the records. A focus group then provided feedback about what was important to retain, features that might add value, aspects of look and feel, and so on. For the rest of December 2002 and into January 2003 detailed planning continued. Data needed to be extracted from the Epublish software and converted to RTF, which would form the basis of the preservation les. The RTF les would be converted to HTML for display via the Departments intranet. XML would be either wrapped around the HTML les, or used to create a separate set of les that, with the metadata, would provide maximum archival content and protection. This would give the greatest exibility to migrate les, while still allowing for lower level browsers that would not support the presentation of data in XML and its associated style sheets. Word was also added as a nal presentation format so that staff using assistive technologies would have the easiest possible access to historical releases. Each of these formats would be stored in a separate directory on a RAID server, using two mirrored hard drives acting as one. After some experimentation, only one reliable way to extract data from the Epublish format was identied to copy and paste the les. Accurate replication of the les was essential because changes in the way the information was presented could affect its interpretation. There were 41 releases to convert this
Case Study 5
217
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
way. Each release had between 35 and 54 chapters, some broken into multiple parts, with at least one table of contents for each chapter, and each of these was a separate le. Files ranged from 6kb to almost 720kb. The associated material for each release would also need to be treated. As the les could only be accessed on a standalone PC, two contractors were employed to begin this task. To convert the RTF les into HTML, a third contractor was engaged to write a prototype. The rst version of the prototype was received in the middle of January 2003. At around the same time, the additional releases found by the software company were delivered, and it was realized that the amount of supplementary materials had been underestimated. The feasibility of a script to automate the Epublish to RTF conversion process was investigated and the script commissioned. Copying and pasting releases continued and the possible overlap of effort was treated as a risk management approach. At the end of January 2003 the prototype HTML format was loaded onto the server, a checklist was created, and a small test was carried out. The results showed that there were no major problems with the design. One issue identied was that the prototype reected the look and feel of later versions of the Guide, rather than earlier ones. The practical outcome was that the original look was reproduced wherever possible, although early releases look far more similar to later ones than they did originally. The conversion script arrived in late February 2003. It was fast and it retained all the important formatting, including bolding, bulleted points, underlining, tables and columns. By comparison, the copy and paste versions had several problems. Early versions of the Guide included coding that was not easily translated to RTF. Underlining, in particular, copied across as coding that replaced the rst and last letter of the underlined segment, and some segments of underlining would not copy at all and had to be retyped. Tabulated data was often columns separated by tabs, and the conversion translated these tabs into single spaces, increasing the difculty of reconstruction. In the end, it was faster to go back to the original Epublish les and run the conversion script over them, and then convert columns to tables. At the same time metadata schemes were examined. The VERS metadata specications were adopted and the metadata was embedded in the head of each master record. XML was investigated in more depth and it was decided to create a separate set of XML les, rather than the XML shell originally considered. This approach provided the exibility needed for display in all browsers and for future migration. By April 2003 the contractors had completed editing the RTF les and were ready to begin the quality control. Attempts were made to rescue data from a number of corrupted les, but none was recovered. The HTML prototype was progressing well, despite the difculties created by the two types of RTF les (scripted and supplied), and the XML and metadata work was well-placed to begin. In May 2003 additional funding was secured to recreate some of the missing releases. Further investigation led us to believe that recreation of most, if not all, of the missing releases was possible, but a signicant amount of work would be required given the lack of a script to automate this process. To recreate a release, the preceding release was copied, and then, guided by the historical update list, the appropriate changes were copied in from a later release, if it existed in an unaltered form, or manually typed in from the paper registry les, if it did not.
Also in May 2003 the XML conversion work commenced. Again, due to the complexity of the Guide, and the lack of available expertise, the XML scripting took longer than anticipated. The style sheets needed to store the data in XML and then display it in XHTML (eXtensible HTML) were complex, because of the required high degree of compliance with the original layout. This resulted in several style sheets (called transformations) used in a cascading fashion to convert the data. FaCS IT staff created an application that would go over the top of the conversion process. This program steps through the processes of taking HTML les, converting them to XHTML, then to XML, and nally cleaning up the nal product. Behind the front end is an off-the-shelf product to convert the HTML les to XHTML, and then another to convert the XHTML to XML, using the customized style sheets. This process was chosen because it will retain the internal hypertext links and formatting. Different style sheets have been created for the different versions of the Guide, and also for the Act. The HTML conversions of the completed releases were also delivered. These les were put through a sampled quality control process designed to pick up as many problems as possible with the least effort. Problems discovered included tables that had not formatted correctly, paragraphs that had failed to split appropriately, and other formatting issues, requiring only minor corrective work. Releases from the most recent years needed far less work because they had been created directly in HTML. Nevertheless, they needed their links checked and broken links modied, especially those to external sites. The releases around the time of the rewrite in 1998 have proved time-consuming. At this time the creators were experimenting with the look of these releases, so the banners and graphics change with every release. In a twelve-month period at the rewrite stage, releases had only been stored as project les in an interim editorial stage rather than as nished HTML, resulting in additional work to publish these releases. Unforeseen issues in this migration process included a major change in the Librarys IT infrastructure which had the unintended consequence of cutting all access to the Guide project data. To solve this, 40 gigabytes of data was transferred to a temporary storage platform during this change. Despite this, the project is close to completion. The HTML version went live on 16 November 2004, and the XML version is expected to be available in 2006 (when the Department rolls out infrastructure able to support XML). The project fullled a clear, identiable and urgent need to preserve data that was becoming corrupted and/or inaccessible. This data is essential corporate history, and there are clear drivers to preserve access to it. Besides the legal responsibility of being able to justify decisions based on data as it was at the time they were made, there is an important need to trace how the management of individual pensions or benets evolved and changed. In some cases signicant changes in meaning have occurred because a word or a bit of formatting has altered. In other cases, understanding the natural progression of changes through time due to legislative and stylistic changes is important to work groups in understanding why a piece of legislation is now interpreted the way that it is. However, at the start of the project the project team had no clear idea of what it was getting into and initial funding for the project was insufcient and more had to be sought. XML was selected with the best of intentions, and it is certainly capable of everything required of it, but lack of experience with it, and the consultants lack of familiarity with the Guide, led to underestimating the
Case Study 6
219
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
amount of work involved in creating the transformations. HTML could have been selected as a reasonable medium for archival purposes, or PDF could have been investigated; these may have ended up being more cost-effective for archiving the material. Using XML means that not all of each complete release is preserved in XML; associated acts and handbooks were not covered due to insufcient funds, and it was decided not to preserve in XML legislation that is not the responsibility of FaCS. This will not change in the foreseeable future.
Case Study 6 PADI (Preserving Access to Digital Information)

This case study describes PADI (Preserving Access to Digital Information), a subject gateway to digital preservation resources It is one of two activities of the National Library of Australia described here, the other being PANDORA (Case Study 2). This case study illustrates the topics of collaboration (Chapter 9) and the importance of education and raising awareness (Chapter 10). PADI (www.nla.gov.au/padi) is a subject gateway to digital preservation resources. PADI is simply described. It was established in 1997 by the National Library of Australia to respond to the growing recognition of the need to safeguard digital heritage. It is intended as an educational tool to share information about and support digital preservation in Australia and worldwide. PADI is now a comprehensive subject gateway to resources dealing with a wide range of digital preservation issues, containing pages about events, policies, strategies, guidelines, and projects, and providing information about organizations, web sites, bibliographies, discussion lists, journals, newsletters, and glossaries. It also offers explanatory text about digital preservation topics. Its signicant amount of explanatory text provides excellent starting points for its users. It is often the only port of call needed for queries about digital preservation matters. PADI also provides other services. One is a moderated discussion list, padiforum-l, for the exchange of news and ideas about digital preservation issues. Its archive is accessible on the web. It also issues an electronic newsletter, DPC/PADI: Whats New in Digital Preservation, in collaboration with the Digital Preservation Coalition. This is a quarterly digest of selected recent digital preservation activity, such as events and publications, compiled from the PADI database, the padiforum-l mailing list, and the digital-preservation mailing list operated by the Digital Preservation Coalition. A feature introduced in 2004 is PADI trails, offering introductions to digital preservation topics by providing guided pathways through materials on the PADI web site. PADI was established in 1997. (Howell and Berthon (2000) give its early history.) Its aims included providing a forum for cross-sectoral cooperation on digital preservation activities, and it is guided by an international advisory group which in 2002 had representatives from the UK, the US, Finland, Norway, Germany, Sweden, Switzerland, Canada, Holland, Norway, as well as Australia. Collaboration is a key characteristic of PADI. Although overall management of PADI is in the hands of the National Library of Australia, it received assistance from CLIR (the US Council on Library and Information Resources) for the Safekeeping Project (see below). The National Library signed a Memorandum of Understanding with the Digital Preservation Coalition (UK) in 2002, resulting
in UK input into PADI and in the joint compilation of the newsletter Whats New in Digital Preservation. ERPANET and the National Library of Australia also signed a Memorandum of Understanding in 2003, to encourage sharing of information resources; this has resulted in increased material from European sources in PADI. A partnership with Nestor, which would bring additional German material to PADI, was being discussed in 2004. Some of the resources on PADI are noted as Safekept. Safekept items are selected as being of long-term value, for example, seminal papers in the eld, or papers that summarize important issues in digital preservation. These are archived and managed by the National Library of Australia in the same way as PANDORA material so that it will be accessible in the long term. The Safekeeping Project is supported by CLIR. Materials are selected for safekeeping by a panel of international selectors. The National Library of Australia contacts the owners of the resources selected to inform them that they will be archiving their resources, and that the owners can reject the offer. This opt-in approach is reported as being 100 per cent successful (Hanley, 2004). PADI is clearly achieving its goals. 97 per cent of a survey of users and stakeholders carried out in 2000 had the PADI site bookmarked in their web browser (Howell and Berthon, 2000, p.12) and it was described by one respondent as a campre around which this community of interest can gather (Howell and Berthon, 2000, p.13). One aspect of the National Library of Australias standard practice in digital preservation appears to be that when a problem is identied, the full range of possible solutions is explored; this proactive process takes place early in the process. PADI exemplies this approach, and also other characteristics of Australian digital preservation practice: sharing of experiences, and a refreshing lack of proprietal ownership. PADI exemplies the cooperative approach that the Library believes will be crucial in securing our global digital memory (Berthon, 2002b).
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Bibliography
The Preface to this book notes that there is a considerable amount of high-quality information available about preserving materials in digital form, and that much of it is available on the web. The accessibility that this provides is countered by the impermanence of much web material, as noted in several chapters in this book. All URLs in this Bibliography were correct at the time of writing. Abbott, D. (2003) Overcoming the dangers of technological obsolescence: rescuing the BBC Domesday Project. DigiCULT.Info: A Newsletter on Digital Culture, 4, 710 Abrams, S.L. and Seaman, D. (2003) Towards a global digital format registry. Paper presented at the 69th IFLA General Conference, Berlin, 19 August 2003. Also available at www.ia.org/IV/ia69/papers/128e-Abrams_Seaman.pdf Access in the future tense. (2004) Washington, D.C.: Council on Library and Information Resources Adcock, E.P. (1998) Principles for the care and handling of library materials. International Preservation Issues, no. 1. Paris: IFLA PAC Adelstein, P.Z. (1993) The stability of optical disks: a science-standards review. The Commission on Preservation and Access Newsletter, 58, 34 Adelstein, P.Z. (2003) Preservation of electronic records: status of ISO standards. In Symposium 2003: Preservation of Electronic Records: New Knowledge and Decision-making, Ottawa, 1518 September 2003: preprints AMOL (2003) Conservation [amol.org.au/craft/conservation/conservation_index.asp] ANICA: Australian Network for Information on Cellulose Acetate (2002) [www.nla.gov. au/anica] Archives Association of Ontario (2001) What to keep and what to destroy? [aao.s. utoronto.ca/aa/appraisal.html] Archiving Web Resources (2004) Themes emerging from Archiving Web Resources Issues for Cultural Heritage Organisations [Conference], National Library of Australia, Canberra, 911 November 2004 [www.nla.gov.au/webarchiving/ConferenceReport. rtf] Arkival Technology Corporation (2002) An evaluation of life expectancies, media stability and performance characteristics of denser magnetic media with special reference to IBM 3590 and digital linear tape systems. Nashua, NH: Arkival Technology Corporation.
222 Bibliography
Arps, M. (1993) CD-ROM: some archival considerations. In Preservation of electronic formats and electronic formats for preservation, ed. J. Mohlhenrich, pp.83107. Fort Atkinson, Wisc: Highsmith Arthur, K. et al. (2004) Recognizing digitization as a preservation reformatting method [www.arl.org/preserv/digit_nal.html] AS ISO 15489.22002 (2002) Records management guidelines. Sydney: Standards Australia Aschenbrenner, A. (2004) The bits and bites of data formats: stainless design for digital endurance. RLG DigiNews, 8 (1) [www.rlg.org/preserv/diginews/v8_n1_feature3. html] Australian Scholarly Editions Centre (2001) Authenticated electronic editions [idun. itsc.adfa.edu.au/ASEC/aueledns.html] Authentec International (2003) Data recovery data loss: case studies index [www.disktape-data-recovery.com/case-studies.htm] Authenticity (1999) [PADI summary]. National Library of Australia [www.nla.gov.au/ padi/topics/4.html] Authenticity in a digital environment. (2000) Washington, D.C.: Council on Library and Information Resources Ayres, M-L. (2001) Evaluating the archives: 20th century Australian literature. Archives & Manuscripts, 29 (2), 3247 Baker, N. (2001) Double fold: libraries and the assault on paper. New York: Random House Barrow, W. J. (1959) Deterioration of book stock: causes and remedies: two studies on the permanence of book paper. Richmond, VA: Virginia State Library Barton, M.R. and Walker, J.H. (2003) Building a business plan for DSpace, MIT Libraries digital institutional repository. Journal of Digital Information, 4 (2) [jodi.ecs.soton.ac.uk/ Articles/v04/i02/Barton] Beagrie, N. (2001) Preserving UK digital library collections. Program, 35 (3), 217226 Beagrie, N. (2003) National digital preservation initiatives: an overview of developments in Australia, France, the Netherlands, and the United Kingdom and of related international activity. Washington, D.C.: Council on Library and Information Resources and Library of Congress Beagrie, N. (2004) The Continuing Access and Digital Preservation Strategy for the UK Joint Information Systems Committee (JISC). D-Lib Magazine, 10 (7/8) [www.dlib.org/ dlib/july04/beagrie/07beagrie.html] Beagrie, N. and Greenstein, D. (1998) A strategic policy framework for creating and preserving digital collections (nal draft, 14/7/98). London: Arts and Humanities Data Service Bearman, D. (1998) Archival methods. Archives and Museums Informatics Technical Report, 9. Pittsburgh: Archives and Records Informatics Bellinger, M. et al. (2004) OCLCs digital preservation program for the next generation library. Advances in Librarianship, 27, 2948 Bennett, J.C. (1997) A framework of data types and formats, and issues affecting the long term preservation of digital material: JISC/NPO studies on the preservation of electronic materials. British Library Research and Innovation Paper, 50. London: British Library Research and Innovation Centre Berthon, H. (2002a) Nurturing our digital memory: digital archiving and preservation at the National Library of Australia [www.nla.gov.au/nla/staffpaper/2002/berthon1. html] Berthon, H. (2002b) Understanding for digital preservation. Gateways, 56 [www.nla. gov.au/ntwkpubs/gw/56/p09a01.htm] Besser, H. (2000) Digital longevity. In Handbook for digital projects: a management tool for preservation and access, ed. M. Sitts. Andover, Mass.: Northeast Document Conservation Center [www.nedcc.org/digital/dighome.htm] Betts, M. (1999) Businesses worry about long-term data losses: will we access our saved data in 20 years? Computerworld News, Sept 20 [www.computerworld.com/news/ 1999/story/0,11280,37036,00.html]
Bibliography
223
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Bishoff, L. and Allen, N. (2004) Business planning for cultural heritage institutions. Washington, D.C.: Council on Library and Information Resources Boadle, D. (2004) Reinventing the archive in a virtual environment: Australians and the non-custodial management of electronic records. Australian Academic & Research Libraries, 35 (3), 242252 Borgman, C.L. (2000) From Gutenberg to the global information infrastructure: access to information in the networked world. Cambridge, Mass.: MIT Press Brabazon, T. (2000) He lies like a rug: digitising memory. Media International Australia, 96, 153161 Bradley, K. (1995) Preservation at the National Library of Australias Oral History Collection. Paper presented at Multimedia Preservation: Capturing the Rainbow Conference, Brisbane, 2830 November 1995 [www.nla.gov.au/nla/staffpaper/npokb.html] Bradley, K. (2000) CD-R archiving: case study of an interim media. Paper presented at A Future for the Past: IASA-SEAPAVAA Conference, Singapore, 37 July 2000. Bradley, K. (2003) Critical choices, critical decisions: sound archiving and changing technology. Paper presented at the Researchers, Communities, Institutions, Sound Recordings Workshop, University of Sydney, 30 September-1 October 2003 [conferences.arts.usyd. edu.au/viewpaper.php?id=57&cf=2] Bradley, K. and Woodyard, D. (2000) Preservation metadata for digital collections: presentation for workshop, Metadata for Long Term Preservation in NEDLIB, Paris, 25 February 2000 [www.nla.gov.au/nla/staffpaper/2000/webb1.html] Breeding, M. (2002) Preserving digital information. Information Today, 19 (5), 4849 Brindley, L.J. (2000) Keynote address to the Preservation 2000 Conference. New Review of Academic Librarianship, 6, 125137 Brown, A. (2003) Preserving the digital heritage: building a digital archive for UK government records. Paper presented at Online Information 2003, 24 December 2003 [www. nationalarchives.gov.uk/preservation/digitalarchive/pdf/brown.pdf] Bryan, R.E. (2003) Survey results: preservation management of born-digital documents in United States manuscripts repositories. In Symposium 2003: Preservation of Electronic Records: New Knowledge and Decision-making, Ottawa, 1518 September 2003: preprints BS4783 (1988- ) Storage, transportation and maintenance of media for use in data processing and information storage. London: British Standards Institution Bullock, A. (1999) Preservation of digital information: issues and current status. Network Notes, 60 [www.collectionscanada.ca/9/1/p1259-e.html] Burrows, T. (2000) Preserving the past, conceptualising the future: research libraries and digital preservation. Australian Academic & Research Libraries, 31 (4), 142153 Byers, F.R. (2003) Care and handling of CDs and DVDs: a guide for librarians and archivists. Washington, D.C.: Council on Library and Information Resources and National Institute of Standards and Technology CAMiLEON (2002?) BBC Domesday [www.si.umich.edu/CAMILEON/domesday/ domesday.html] Caplan, P. (2004) PREMIS Preservation Metadata Implementation Strategies update 1. Implementing preservation repositories for digital materials: current practice and emerging trends in the cultural heritage community. RLG DigiNews, 8 (5) [www.rlg. org/en/page.php?Page_ID=20462#article2] Casson, L. (2002) Libraries in the ancient world. New Haven: Yale Nota Bene Cedars Project Team (2002) The Cedars Project Report, April 1998-March 2001 [www. leeds.ac.uk/cedars/OurPublications/cedarsrepmar01exec.html] Chapman, S. (2003) Counting the cost of digital preservation: Is repository storage affordable? Journal of Digital Information, 4 (2), 115 [jodi.ecs.soton.ac.uk/Articles/v04/i02/ Chapman] Chivers, A. and Feather, J. (1998) The management of digital data: a metadata approach. Electronic Library, 16 (6), 365372
224 Bibliography
Christensen, N.H. (2004) Towards format repositories for web archives. Paper presented at the 4th International Web Archiving Workshop (IWAW04), Bath, 16 September 2004. Available at www.netarchive.dk/website/publications/FormatRepositories-2004.pdf Clausen, L.R. (2004) Handling le formats. rhus: State and University Library Cloonan, M.V. (1993) The preservation of knowledge. Library Trends, 41 (4), 594605 Cloonan, M.V. (2001) W(h)ither preservation? Library Quarterly, 71 (2), 231242 Cloonan, M.V. and Sanett, S. (2000) Comparing preservation strategies and practices for electronic records. New Review of Academic Librarianship, 6, 205216 Cloonan, M.V. and Sanett, S. (2002) Preservation strategies for electronic records: where are we now obliquity and squint? American Archivist, 65, 70106 Condron, F. et al. (1999) Strategies for digital data: ndings and recommendations from digital data in archaeology: a survey of user needs. York: Archaeology Data Service [ads.ahds.ac. uk/project/strategies/index.html] Consultative Committee for Space Data Systems (2002) Reference Model for an Open Archival Information System (OAIS). CCSDS 650.0-B-1 Blue Book. Washington, D.C.: NASA Conway, P. (1996) Preservation in the digital world. Washington, D.C.: Council on Library and Information Resources Conway, P. (2000) Overview: rationale for digitization and preservation. In Handbook for digital projects: a management tool for preservation and access, ed. M. Sitts. Andover, Mass.: Northeast Document Conservation Center [www.nedcc.org/digital/dighome.htm] Cook, T. (1995) Its 10 oclock: do you know where your data are? Technology Review January [web.mit.edu/erm/tcook/tr1995.html] Cook, T. (2000) Beyond the screen: the records continuum and archival cultural heritage. Paper presented at the Australian Society of Archivist Conference, Melbourne, 18 August 2000 [www.archivists.org.au/sem/conf2000/terrycook.pdf] Cornwell Management Consultants (2004) Training needs analysis: nal report. London: Digital Preservation Coalition Council of Australian State Libraries (2002) Background paper [for] the Regional Consultation on the Preservation of Digital Heritage for Asia and the Pacic, National Library of Australia, Canberra, 46 November, 2002 Cox, R.J. (2000?) The functional requirements for evidence in recordkeeping [www2.sis. pitt.edu/~rcox/FunReqs.htm] Cox, R.J. (2001) Managing records as evidence and information. Westport, Conn.: Quorum Books Cox, R.J. (2002) Vandals in the stacks: a response to Nicholson Bakers assault on libraries. Westport, Conn: Greenwood Press Dack, D. (2001) Persistence is a virtue. Paper presented at the Information Online 2001 Conference, Sydney, 1618 January 2001 [www.nla.gov.au/nla/staffpaper/2001/dack. html] Dale, R.L. (2004) Consortial actions and collaborative achievements: RLGs preservation program. Advances in Librarianship, 27, 123 Darlington, J., Finney, A. and Pearce, A. (2003) Domesday Redux: the rescue of the BBC Domesday Project videodiscs. Ariadne, 36 [www.ariadne.ac.uk/issue36/tna] Day, M. (2004) Preservation metadata. In Metadata Applications and Management ed. G.E. Gorman and D.G. Dorner, pp.253273. London: Facet Day, R. (1989) Wheres the rot?: a special report on CD longevity. Stereo Review, April, 2324. de Lusenet, Y. (2002) Preservation of digital heritage: draft discussion paper prepared for UNESCO. European Commission on Preservation and Access [www.knaw.nl/ecpa/ PUBL/unesco.html] Deegan, M. and Tanner, S. (2002) The digital dark ages. Update, May [www.cilip.org. uk/publications/updatemagazine/archive/archive2002/may/update0205b.htm] Digital Curation Centre (2004) What is digital curation? [www.dcc.ac.uk/what.html]
Bibliography
225
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Digital Domesday Book unlocked. (2002) BBC News, 2 December [news.bbc.co.uk/1/hi/ technology/2534391.stm] Digital Preservation Coalition (2003) Annual company report 23 July 200231 July 2003 [www.dpconline.org/docs/DPCAR0203.pdf] Digital Preservation Coalition (2004) Annual company report 1 August 200331 July 2004 [www.dpconline.org/docs/DPCAR0304.pdf] Digital Preservation Coalition (2004) Digital preservation: the global context: report on the DPC Forum held at the British Library Conference Centre, Wednesday 23 June 2004 [www.dpconline.org/graphics/events/for040623.html] Digital Preservation Testbed (2003) From Digital Volatility to Digital Permanence: Preserving Email. The Hague: ICTU [www.digitaleduurzaamheid.nl/bibliotheek/docs/volatilitypermanence-email-en.pdf] DSpace communications kit. (2003) [dspace.org/implement/communications-kit.html] DSpace FAQ. (2003) [dspace.org/faqs/index.html] Dunning, A. (2001) Excavating data: retrieving the Newham Archive. London: Arts and Humanities Data Service [ahds.ac.uk/creating/case-studies/newham] Duranti, L. (2003) Authenticity requirements for electronic records. In Symposium 2003: Preservation of Electronic Records: New Knowledge and Decision-making, Ottawa, 1518 September 2003: preprints Dureau, J.M. and Clements, D.W.G. (1986) Principles for the preservation and conservation of library materials. IFLA Professional Reports, no. 8. The Hague: IFLA Eastwood, T. (2002) The appraisal of electronic records: what is new? Comma, 20021/2, 7787 Eastwood, T. (2003) Appraising digital material for preservation as cultural heritage. In Symposium 2003: Preservation of Electronic Records: New Knowledge and Decision-making, Ottawa, 1518 September 2003: preprints Edmondson, R. (1998) A philosophy of audiovisual archiving. Paris: UNESCO [unesdoc. unesco.org/images/0011/001131/113127eo.pdf] Edmondson, R. (2002) Memory of the World: general guidelines rev. edn. Paris: UNESCO Edmondson, R. (2004) Audiovisual archiving: philosophy and principles. Paris: UNESCO [unesdoc.unesco.org/images/0013/001364/136477e.pdf] Electronic Literature Organization (2005) [www.eliterature.org] Ellis, J. (1993) (ed.) Keeping archives 2nd edn. Melbourne: Thorpe in association with the Australian Society of Archivists Ellis, S. (2003) The moving frontier: the preservation of authentic and reliable digital records over time. In GLAM: galleries, libraries, archives and museums: conference proceedings, Australian Society of Archivists 2003 National Conference, 1720 September 2003, Adelaide, pp.121125. OConnor, A.C.T.: Australian Society of Archivists Ellis, S. and Stuckey, S. (1994) Australian Archives approach to preserving long-term access to the Commonwealths electronic records. Playing for keeps the proceedings of an electronic records management conference, pp.113132. Canberra: Australian Archives ERPANET (2003) Selecting technologies tool [www.erpanet.org/guidance/docs/ ERPANETSelect_Techno.pdf] ERPANET/CODATA Workshop (2004) The selection, appraisal and retention of digital scientic data: nal report, ERPANET/CODATA Workshop, Biblioteca Nacional, Lisbon, December 1517, 2003. [www.erpanet.org/events/2003/lisbon/LisbonReport Final.pdf] ERPATraining (2004) File formats for preservation: brieng paper [www.erpanet.org/ events/2004/vienna/erpaTrainingWien_BriengPaper_v02.pdf] Exon, M. (1995) The long-term management issues in the preservation of electronic information. Paper presented at Multimedia Preservation: Capturing the Rainbow Conference, Brisbane, 2830 November 1995 [www.nla.gov.au/niac/meetings/npo95me.html] Farley, J. (1999) An introduction to archival materials: new media. Kew, Surrey: Public Record Ofce.
226 Bibliography
Faulds, F. and Challinor, A. (1998) Archiving from the data warehouse. Information Management & Technology, 31 (6), 278280 Flesch, J. (1996) A labour of love?: the story behind the compilation of Love brought to book: a bio-bibliography of 20th century Australian romance novels. Australian Academic &Research Libraries, 27 (3), 182190 Flesch, J. (2004) From Australia with love: a history of modern Australian popular romance novels. Fremantle: Curtin University Books. Gatenby, P. (2002a) Legal deposit, electronic publications and digital archiving: The National Library of Australias Experience. Paper presented at the 68th IFLA General Conference, Glasgow, August 1824, 2002 [www.ia.org/IV/ia68/papers/071124e. pdf] Gatenby, P. (2002b) Report on Senior Executive Fellowship to research digital archiving in national libraries [www.nla.gov.au/nla/staffpaper/2002/elect.html] George, J. (2003) What users are telling us: a symposium. CLIR Issues, 33, 12, 6 Gertz, J. (1998) Selection guidelines for preservation. Paper presented at the Joint RLG and NPO Preservation Conference: Guidelines for Digital Imaging, University of Warwick, 2830 September 1998 [www.rlg.org/preserv/joint/gertz.html] Gertz, J. (2000) Selection for preservation in the digital age. Library Resources & Technical Services, 44 (2), 97104 Geser, G. and Pereira, J. (2004) (eds) The future digital heritage space: an expedition report. DigiCULT Thematic Issue, 7 Gilliland-Swetland, A.J. (2000) Enduring paradigm: the value of the archival perspective in the digital environment. Washington, D.C.: Council on Library and Information Resources. Gilliland-Swetland, A.J. (2002) Testing our truths: delineating the parameters of the authentic archival electronic record. American Archivist, 65, 196215 Gilliland-Swetland, A.J. (2005) Electronic records management. ARIST: Annual Review of Information Science and Technology, 39, 219253 Global Data Vault Inc. (2005) dataVault service overview [www.globaldatavault.com/v3/ howitworks/serviceoverview.asp] Gorman, M. (1997) What is the future of cataloguing and cataloguers? Paper presented at the 63rd IFLA General Conference, Copenhagen, Denmark, August 31-September 5, 1997 [www.ia.org/IV/ia63/63gorm.htm] Gracy, K.F. (2001) The imperative to preserve: competing denitions of value in the world of lm preservation. PhD Thesis, University of California Los Angeles, USA Granger, S. (2000) Emulation as a digital preservation strategy. D-Lib Magazine, 6 (10) [www.dlib.org/dlib/october00/granger/10granger.html] Green, A., Dionne, J. and Dennis, M. (1999) Preserving the whole: a two-track approach to rescuing social science data and metadata. Washington, DC: Council on Library and Information Resources Greenan, M. (2003) Dspace. DigiCULT.Info: A Newsletter on Digital Culture, 3, 8 Guenther, R. (2004) PREMIS: Preservation Metadata Implementation Strategies update 2: core elements for metadata to support digital preservation. RLG DigiNews, 8 (6) [www.rlg.org/en/page.php?Page_ID=20492#article2] Gunton, T. (1993) A dictionary of information technology and computer science 2nd edn. Manchester: NCC Blackwell Gwinn, N.E. (1993) A national preservation program for agricultural literature. [preserve. nal.usda.gov:8300/npp/presplan.htm] Hafner, K. (2004) Even digital memories fade. New York Times, 10 November 2004. Hanley, M. (2004) PADI (Preserving Access to Digital Information) and safekeeping High Energy Physics Libraries Webzine, 9. Also available at www.nla.gov.au/nla/staffpaper/ 2004/hanley1.html Harris, C. (2000) Selection for preservation. In Preservation: issues and planning, ed. P.N. Banks and R. Pilette, pp.206224. Chicago: American Library Association. Harter, R. (1999) Piltdown Man [home.tiac.net/~cri_a/piltdown/piltdown.html]
Bibliography
227
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Harvey, R. (1993) Preservation in libraries: principles, strategies and practices for librarians. London: Bowker-Saur Harvey, R. (1995) From digital artefact to digital object. Paper presented at Multimedia Preservation: Capturing the Rainbow Conference, Brisbane, 2830 November 1995 [www. nla.gov.au/niac/meetings/npo95rh.html] Harvey, R. (2003) Preserving digital documentary heritage in libraries: what do we select? In Symposium 2003: Preservation of Electronic Records: New Knowledge and Decision-making, Ottawa, 1518 September 2003: preprints Heazlewood, J. (2000) ePreservation in the archive: theories, practices. Australian Academic & Research Libraries, 31 (4), 173187 Heazlewood, J. (2002) Public Record Ofce Victoria: background report to the UNESCO Regional Consultation on the Preservation of Digital Heritage, Canberra, 46 November 2002. Hebditch, M. (1998) Foreword. In Orna, E. and Pettitt, C. Information Management in Museums 2nd edn, p.viii. Aldershot: Gower Hedstrom, M. (1998) Digital preservation: a time bomb for digital libraries. Computers and Humanities, 31, 189202 Hedstrom, M. and Lampe, C. (2001) Emulation vs. migration: do users care? RLG DigiNews, 5 (6) [www.rlg.org/legacy/preserv/diginews/diginews56.html#feature1] Hedstrom, M. and Lee, C.A. (2002) Signicant properties of digital objects: denitions, applications, implications. Proceedings of the DLM-Forum 2002: @ccess and Preservation of Electronic Information: Best Practices and Solutions, Barcelona, 68 May 2002, pp. 218223. Brussels: European Commission Hedstrom, M. and Montgomery, S. (1999) Digital preservation needs and requirements in RLG member institutions: a study commissioned by the Research Libraries Group. Mountain View, CA: Research Libraries Group Hendley, T. (1998) Comparison of methods & costs of digital preservation. London: British Library Research and Innovation Centre Heslop, H., Davis, S. and Wilson, A. (2002) An approach to the preservation of digital records. Canberra: National Archives of Australia. Heterick, B. (2002) Applying the lessons learned from retrospective archiving to the digital archiving conundrum. Information Services & Use, 22, 113120 Higgs, E. (1998) (ed.) History and electronic artefacts. Oxford: Clarendon Press Hodge, G. and Frangakis, E. (2004) Digital preservation and permanent access to scientic information: the state of the practice: a report sponsored by the International Council for Scientic and Technical Information (ICSTI) and CENDI [cendi.dtic.mil/ publications/043dig_preserv.html] Hodge, G.M. (2003) Digital preservation: overview of current developments. Information Services & Use, 22, 7382 Hofman, H. (2002) Review: some comments on preservation metadata and the OAIS Model. DigiCULT.Info: A Newsletter on Digital Culture, 2, 1520 Holdsworth, D. and Wheatley, P. (2001) Emulation, preservation, and abstraction. RLG DigiNews, 5 (4) [www.rlg.org/preserv/diginews/diginews54.html] Horn, A. (2002) [Background paper on Council of Australian University Librarians for] The Regional Consultation Meeting on the Preservation of Digital Heritage for Asia and the Pacic, National Library of Australia, Canberra, 46 November 2002 [www.anu.edu.au/caul/preservation/DigitalHeritage-CAUL.doc] Howell, A. (2000) Perfect one day digital the next: challenges in preserving digital information. Australian Academic & Research Libraries, 31 (4), 121141 Howell, A. (2001) Preserving information in a digital age: whats the difference? Paper Conservator, 25, 133149 Howell, A.G. (2004) Preserving digital information: challenges and solutions: workbook [www.alanhowell.com.au/dpw/pdi_wkb.pdf] Howell, A. and Berthon, H. (2000) http://www.nla.gov.au/padi/: Preserving Access to Digital Information (PADI) an opportunity for global cooperation. Paper presented
228 Bibliography
at IFLA Symposium on Managing the Preservation of Periodicals and Newspapers, Paris, 2124 August 2000 [www.ia.org.sg/VI/4/conf/howell.pdf] Humphrey, C. (2003) Preserving research data: a time for action. In Symposium 2003: Preservation of Electronic Records: New Knowledge and Decision-making, Ottawa, Canada, 1518 September 2003: Preprints. Inera Inc. (2001) E-journal archive DTD feasibility study prepared for the Harvard University E-Journal Archiving Project. Newton, Mass. [www.diglib.org/preserve/hadtdfs.pdf] Integrity and authenticity of digital cultural heritage objects (2002) DigiCULT Thematic Issue, 1. Salzburg: DigiCULT Project [www.digicult.info/downloads/thematic_issue_1_nal. pdf] International Association of Sound and Audiovisual Archives Technical Committee (2004) Guidelines on the production and preservation of digital audio objects. Aarhus: IASA. International Council on Archives Committee on Electronic Records (1997) Guide for managing electronic records from an archival perspective. ICA Studies, 8. Paris: ICA International DOI Foundation (2004?) Welcome to the digital object identier system [www.doi.org] James, H. et al. (2003) Feasibility and requirements study on preservation of e-prints: report commissioned by the Joint Information Systems Committee. London: JISC [www.jisc. ac.uk/uploaded_documents/e-prints_report_nal.pdf] James, M.S. (2001) Fading bits of history as computer records replace paper. Good Morning America World News Tonight, 9 July 2001. Available at www.thearchivecompany.com/ Forms/Fading%20bits%20of%20history.htm Janes, J. (2003) Internet librarian: authority by community. American Libraries, January, 92 Jenkinson, H. (1965) A manual of archive administration 2nd edn rev. London: Lund, Humphries JISC (2004) Digital preservation training programme [www.jisc.ac.uk/index.cfm?name= project_dptp] Jones, M. (2004) Editors interview: Digital Preservation Coalition. RLG DigiNews, 8 (1) [www.rlg.org/preserv/diginews/diginews81.html] Jones, M. and Beagrie, N. (2001) Preservation management of digital materials: a handbook. London: British Library. Kasteren, J. van (2002) Analogue versus digital storage of our history: an interview with Hartmut Weber, President, Bundesarchiv, Germany. DigiCULT Thematic Issue, 1, 9 [www.digicult.info/downloads/thematic_issue_1_nal.pdf] The KB e-Depot: Permanent Access to Digital Publications (2004) [publicity brochure] Kenney, A.R. and Rieger, O.Y. (2000) Moving theory into practice: digital imaging for libraries and archives. Mountain View, CA: Research Libraries Group Kenney, A.R and Rieger, O.Y. (2001) The National Library of Australias Digital Preservation Agenda: an interview with Colin Webb. RLG DigiNews, 5 (1) [www.rlg.org/ legacy/preserv/diginews/diginews51.html#feature1] Kenney, A.R. and Stam, D.C. (2002) The state of preservation programs in American college and research libraries: building a common understanding and action agenda. Washington, D.C.: Council on Library and Information Resources Kenney, A. et al. (2003) Digital preservation management: implementing short-term strategies for long-term problems [tutorial]. Ithaca, N.Y.: Cornell University Library [www.library. cornell.edu/iris/tutorial/dpm/index.html] Kerry, A. (2001) Digital preservation: the research and development agenda in Australia. Paper prepared for the Digital Continuity Forum, Swinburne University of Technology, 19 November 2001. Melbourne: Australian Foresight and Information Resources Group, Swinburne University of Technology Knight, S. (2003) Preservation metadata and digital continuity. DigiCULT.Info: A Newsletter on Digital Culture, 3, 1820 Koehler, W. (2004) A longitudinal study of web pages continued: a consideration of document persistence. Information Research 9 (2) [informationr.net/ir/92/paper174.html]
Bibliography
229
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Koerbin, P. (2004) Managing web archiving in Australia: a case study. Paper presented at the 4th International Web Archiving Workshop (IWAW04), Bath, 16 September 2004 [www.iwaw.net/04] Kranch, D.A. (1998) Beyond migration: preserving electronic documents with digital tablets. Information Technology and Libraries, 17 (3), 138146 Kunej, D. (2004) The Use of CD-Rs for archival storage and preservation: vulnerability of CD-Rs based on the susceptibility to sunlight experiment. IASA Journal, 23, 6580 Kuny, T. (1998) The digital dark ages?: challenges in the preservation of electronic information. International Preservation News, 17, 813 Lacken, C. (2001) Prioritising for migration to digital formats in television archives. IASA Journal, 17, 2528. Lavoie, B. (2000) Meeting the challenge of digital preservation: the OAIS Reference Model. OCLC Newsletter, Jan/Feb, 2630 Lavoie, B.F. (2003) The incentives to preserve digital materials: roles, scenarios, and economic decision-making. Dublin, Ohio: OCLC [www.oclc.org/research/projects/digipres/ incentives-dp.pdf] Lavoie, B.F. (2004) Of mice and men: economically sustainable preservation for the twenty-rst century. In Access in the future tense (2004), pp.4554 Lavoie, B. and Dempsey, L. (2004) Thirteen ways of looking at . . . digital preservation. D-Lib Magazine, 10 (7/8) [www.dlib.org/dlib/july04/lavoie/07lavoie.html] Lawrence, G.W. et al. (2000) Risk management of digital information: a le format investigation. Washington, D.C.: Council on Library and Information Resources Lazinger, S.S. (2001) Digital preservation and metadata: history, theory, practice. Englewood, Colo: Libraries Unlimited LeFurgy, W.G. (2003) PDF/A: developing a le format for long-term preservation. RLG DigiNews, 7 (6) [www.rlg.org/legacy/preserv/diginews/diginews76.html#feature1] Legal deposit (2003) [PADI summary]. National Library of Australia [www.nla.gov.au/ padi/topics/67.html] Lieberman, P. (1995) Taking measure of magnetic, optical, and magneto-optical media and drives, CD-ROM Professional, 8 (7), 6274 Life Caching (2004) [www.trendwatching.com/trends/LIFE_CACHING.htm] Lim S. L., Ramaiah, C. K. and Pitt K.W. (2003) Problems in the preservation of electronic records. Library Review, 52 (3), 117125 Linden, J. et al. (2005) The Large-Scale Archival Storage of Digital Objects: Technology Watch Report. DPC Technology Watch Series Report, 0403. York: Digital Preservation Coalition [www.dpconline.org/docs/dpctw0403.pdf] LOCKSS (2004?) About LOCKSS [lockss.stanford.edu/about/about.htm] Lord, P. and Macdonald, A. (2003) E-science curation report: data curation for e-science in the UK: an audit to establish requirements for future curation and provision. Prepared for the JISC Committee for the Support of Research (JCSR). London: Digital Archival Consultancy Lorie, R. (2002) The UVC: a method of preserving digital documents proof of concept. Longterm preservation study report series, no. 4. Amsterdam: IBM Netherlands Lowenthal, D. (1985) The past is a foreign country. Cambridge: Cambridge University Press Lukesh, S.S. (1999) E-mail and potential loss to future archives and scholarship or the dog that didnt bark. Firstmonday, 4 (9) [rstmonday.org/issues/issue4_9/lukesh/ index.html] Lupovici, C. (2001) Technical data and preservation needs. Paper presented at the 67th IFLA Council and General Conference, Boston, August 1625, 2001 [www.ia.org/IV/ ia67/papers/163168e.pdf] Lyall, J. (1993) Determining the signicance of documentary heritage materials. Australian Academic & Research Libraries, 24, 6977 Lyman, P. and Kahle, B. (1998) Archiving digital cultural artifacts: organizing an agenda for action. D-Lib Magazine, July/August [www.dlib.org/dlib/july98/07lyman.html]
230 Bibliography
Lynch, C. (1999) Canonicalization: A fundamental tool to facilitate preservation and management of digital information. D-Lib Magazine, 5 (9) [www.dlib.org/dlib/ september99/09lynch.html] Lynch, C.A. (2004) Editors interview with Clifford A. Lynch. RLG DigiNews, 8 (4) [www.rlg.org/en/page.php?Page_ID=19481#article0] Marcum, D. and Friedlander, A. (2003) Keepers of the crumbling culture: what digital preservation can learn from library history. D-Lib Magazine, 9 (5) [www.dlib.org/ dlib/may03/friedlander/05friedlander.html] McKnight, S. and Livingston, H. (2003) Reuse of learning objects? why, how and when? Educause in Australasia, University of Adelaide, 69 May 2003 [www.deakin.edu.au/ learningservices/staff/publications/sue/ReuseLearningObjects.doc] Meeting of Experts on Digital Preservation (2004) Report on the Meeting of Experts on Digital Preservation: Metadata Specications. Washington, D.C.: USGPO [www.gpoaccess.gov/about/reports/metadata.html] Mellor, P. (2003) CAMiLEON: Emulation and BBC Domesday, RLG DigiNews, 7 (2) [www.rlg.org/legacy/preserv/diginews/diginews72.html#feature3] Millar, L. (2004) Authenticity of electronic records: a report prepared for UNESCO and the International Council on Archives. ICA Study, 132. Paris: ICA. Moehle, F. (2004) The role of le formats in digital preservation: opportunities and threats. Paper presented at the ERPANET File Formats for Preservation seminar, Vienna, 1011 May 2004 [www.erpanet.org/events/2004/vienna/index.php#papers] Montfort, N. and Wardrip-Fruin, N. (2004) Acid-free bits: recommendations for long-lasting electronic literature. Los Angeles, CA.: Electronic Literature Organization [www.eliterature.org/pad/afb.html] Morris, S. (2002) The preservation problem: collaborative approaches. Information Services & Use, 22, 127132 Morrison, I. (1999) www.nla.gov.au/pandora: Australias internet archive. Australian Library Journal, 48 (3), 271284 Mller, E. (2004) DiVA: academic archive online. DigiCULT.Info: A Newsletter on Digital Culture, 9, 2225 Muir, A. (2004a) Digital preservation: awareness, responsibility and rights issues. Journal of Information Science, 30 (1), 7392. Muir, A. (2004b) Issues in the long-term management of digital materials. In Managing preservation for libraries and archives: current practice and future developments, ed. J. Feather, pp.6781. Aldershot, Hants: Ashgate MyLifeBits Project (2003) [research.microsoft.com/barc/MediaPresence/MyLifeBits. aspx] National Archives (U.K.) (2004?) 1986 Domesday Community [www.nationalarchives.gov.u k/preservation/research/domesday.htm] National Archives of Australia (2001a) Archiving web resources: a policy for keeping records of web-based activity in the Commonwealth Government. Canberra: National Archives of Australia National Archives of Australia (2001b) Archiving web resources: guidelines for keeping records of web-based activity in the Commonwealth Government. Canberra: National Archives of Australia National Archives of Australia (2001c) Recordkeeping metadata standard for Commonwealth agencies [www.naa.gov.au/recordkeeping/control/rkms/summary.htm] National Archives of Australia (2004) Digital Preservation Project (previously known as the Agency to Researcher (AtoR) project) [www.naa.gov.au/recordkeeping/preservation/digital/summary.html] National Library of Australia (1999?) Voluntary deposit scheme for physical format electronic publications [www.nla.gov.au/policy/vdelec.html] National Library of Australia (1999a) A draft research agenda for the preservation of physical format digital publications [www.nla.gov.au/policy/rsagenda.html]
Bibliography
231
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
National Library of Australia (1999b) Preservation metadata for digital collections [www.nla.gov.au/preserve/pmeta.html] National Library of Australia (2000?) First steps in preserving digital publications [www.nla.gov.au/pres/epupam.html] National Library of Australia (2001) Use of Australian CD-ROMs & other electronic materials acquired by deposit, version 2 [www.nla.gov.au/policy/cdrom.html] National Library of Australia (2002a) Digital preservation policy [www.nla.gov.au/ policy/digpres.html] National Library of Australia (2002b) Persistent identiers [www.nla.gov.au/initiatives/persistence.html] National Library of Australia (2004?) Recommended practices for digital preservation [www.nla.gov.au/preserve/digipres/digiprespractices.html] National Research Council (1995) Preserving scientic data on our physical universe: a new strategy for archiving the nations scientic information resources. Washington, D.C.: National Academy Press NISO (2004) Understanding metadata. Bethesda, MD: National Information Standards Organization Press Norsam Technologies (2001) HD-Rosetta Archival Preservation Services [www.norsam.com/ hdrosetta.htm] NSF-DELOS Working Group on Digital Archiving and Preservation (2003) Invest to save: report and recommendations of the NSF-DELOS Working Group on Digital Archiving and Preservation. National Science Foundation & The European Union Nunberg, G. (1995) The places of books in the age of electronic reproduction. In Future libraries, ed. R.H. Bloch and C.A. Hesse, pp.1337. Berkeley: University of California Press OCLC/RLG Working Group on Preservation Metadata (2002) A metadata framework to support the preservation of digital objects [www.oclc.org/research/projects/pmwg/ pm_framework.pdf] OCLC/RLG PREMIS Working Group (2004) Implementing preservation strategies for digital materials: current practice and emerging trends in the cultural heritage community. Dublin, OH: OCLC OMahony, D.P. (1998) Here today, gone tomorrow: what can be done to assure permanent public access to electronic government information? Advances in Librarianship, 22, 107121 Ontrack Data Recovery (2004) Understanding data loss [www.ontrack.com/understandingdataloss/index.asp] Pacey, A. (1991) Developing selection criteria for special collections. Canadian Library Journal, 48, 187190 Parkes, M. (1999) A review of the preservation issues associated with digital documents. Australian Library Journal, 48 (4), 358377 Parliament of Australia Joint Committee of Public Accounts and Audit (2004) Report 399: inquiry into the management and integrity of electronic information in the Commonwealth, March 2004. Canberra: Parliament of Australia Pawley, A. (2003) The need for a Pacic languages archive. Paper presented at the Researchers, Communities, Institutions, Sound Recordings Workshop, 30 September-1 October 2003 [conferences.arts.usyd.edu.au/viewpaper.php?id=45&cf=2] Peasley, J. (1993) Australian-created electronic information: preservation issues and guidelines. Australian Library Journal, 42, 147141 Persistent identiers (2002) [PADI summary]. National Library of Australia [www.nla. gov.au/padi/topics/36.html] Phillips, M. (2000) Managing chaos in the cyberworld. Paper presented at CONSAL 2000, the 11th Congress of Southeast Asian Librarians Conference, Singapore, 2628 April 2000 [www.nla.gov.au/nla/staffpaper/mphillips5.html]
232 Bibliography
Phillips, M. (2002) Archiving the Web: the National Collection of Australian Online Publications. Paper presented at the International Symposium on Web Archiving, National Diet Library, Tokyo, Japan, 30 January 2002. [www.nla.gov.au/nla/staffpaper/2002/phillips1.html] Phillips, M. (2004) What to collect and how to do it: the National Library of Australias selective approach. Paper presented at the Archiving Web Resources International Conference, National Library of Australia, Canberra, 911 November 2004 [www.nla. gov.au/webarchiving/PhillipsMargaret.rtf] Piggott, M. (2001) Appraisal: the state of the art: paper delivered at a professional development workshop presented by ASA South Australia Branch 26 March 2001 [www. archivists.org.au/sem/misc/piggott.html] Pockley, S. (1995 ) Killing the duck to keep the quack [www.duckdigital.net/FOD/ FOD0055.html] Pockley, S. (2002) Australian Centre for the Moving Image: background report to the UNESCO Regional Consultation on the Preservation of Digital Heritage, Canberra, 46 November 2002. Potter, M. (2002) Researching long term digital preservation approaches in the Dutch Digital Preservation Testbed (Testbed Digital Bewaring). RLG DigiNews, 6 (3) [www. rlg.org/legacy/preserv/diginews/diginews63.html#feature2] Preservation metadata (2003) [PADI summary]. National Library of Australia [www.nla.gov.au/padi/topics/32.html] Public Record Ofce (2000) Evaluating information assets: appraising the inventory of electronic records. Kew, Public Record Ofce [www.nationalarchives.gov.uk/ electronicrecords/advice/pdf/appraisal_toolkit.pdf] Public Record Ofce of Victoria (1996) Keeping records forever: records management vision development. Melbourne, Vic. [www.prov.vic.gov.au/vers/pdf/kerf.pdf] Public Record Ofce of Victoria (1998) Victorian Electronic Records Strategy nal report. Melbourne, Vic. [www.prov.vic.gov.au/vers/pdf/nal.pdf] Public Record Ofce of Victoria (2003) Management of electronic records PROS 99/007 (Version 2). Melbourne, Vic. [www.prov.vic.gov.au/vers/standard/version2.htm] Public Record Ofce of Victoria (2004) VERS: Forever Digital. Melbourne, Vic. Quenault, H. (2001) VERS implementation project at the Department of Infrastructure, Melbourne, Australia. Records Management Journal, 11 (2), 7182 Quenault, H. (2003) Case study: implementation of the VERS strategy for electronic records within the Government of the State of Victoria, Australia. Paper presented at Symposium 2003: Preservation of Electronic Records: New Knowledge and Decision-making, Ottawa, Canada, 1518 September 2003 Quenault, H. (2004) VERS: building a digital record heritage. Paper presented at VALA 2004 12th Biennial Conference and Exhibition, Melbourne, 35 February, 2004 [www. vala.org.au/vala2004/2004pdfs/13Quena.PDF] Read, J. (1999) Parish map preservation project: the use of digital holography, CDs [sic] and computers to copy, view (and preserve) old maps. LASIE, 30 (3), 4553 Redening lm preservation: a national plan (1994) Washington, D.C.: Library of Congress Reich, V.A. (2002) Lots of Copies Keep Stuff Safe as a cooperative archiving solution for e-journals. Issues in Science and Technology Librarianship, 36 [www.istl.org/02-fall/ article1.html] Reich, V. and Rosenthal, D.S. (2000) LOCKSS (Lots Of Copies Keep Stuff Safe). New Review of Academic Librarianship, 6, 155161 Reuben, L. (2004) From proprietary format to open source preserving digital resources. Paper presented at the Digital Library Technology Conference, Sydney, July 2004. RLG/OCLC Working Group on Digital Archive Attributes (2002) Trusted digital repositories: attributes and responsibilities. Mountain View, CA: Research Libraries Group Robertson, A. and Cunningham, A. (2000) Documenting the business of government: archival issues in the digital age. Australian Academic & Research Libraries, 31 (4), 188201
Bibliography
233
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Rosenthal, D.S. et al. (2005) Transparent format migration of preserved web content. DLib Magazine, 11 (1) [www.dlib.org/dlib/january05/rosenthal/01rosenthal.html] Ross, S. (2000) Changing trains at Wigan: digital preservation and the future of scholarship. London: National Preservation Ofce, British Library [www.bl.uk/services/preservation/occpaper.pdf] Ross, S. (2002) Position paper on integrity and authenticity of digital cultural heritage objects. DigiCULT Thematic Issue, 1, 78 [www.digicult.info/downloads/thematic_ issue_1_nal.pdf] Ross, S. (2004) The role of ERPANET in supporting digital curation and preservation in Europe. D-Lib Magazine, 10 (7/8) [www.dlib.org/dlib/july04/ross/07ross.html] Ross, S. and Gow, A. (1999) Digital archaeology: rescuing neglected and damaged data Resources: a JISC/NPO study within Electronic Libraries (eLib) Programme on the Preservation of Electronic Materials. London: Library Information Technology Centre [www.ukoln. ac.uk/services/elib/papers/supporting/pdf/p2.pdf] Ross, S., Greenan, M. and McKinney, P. (2003) Cross-sectoral development of digital preservation strategies: ERPANET and the expansion of knowledge. In Symposium 2003: Preservation of Electronic Records: New Knowledge and Decision-making, Ottawa, 1518 September 2003: preprints Ross, S. et al. (2004) New organisational structures responding to new challenges: the Digital Curation Centre in the UK. DCC Public Lecture, Schweizerisches Bundesarchive, 25 October 2004 [www.erpanet.org/events/2004/bern/SR_DCCpresentation_ BERNE_erpanet_mtg_2.pdf] Rothenberg, J. (1995) Ensuring the longevity of digital documents. Scientic American, 272, 4247 Rothenberg, J. (1999a) Avoiding technological quicksand: nding a viable technical foundation for digital preservation. Washington, D.C.: Council on Library and Information Resources Rothenberg, J. (1999b) Ensuring the longevity of digital information. Washington, D.C.: Council on Library and Information Resources (Expanded version of Rothenberg 1995) [www.clir.org/pubs/archives/ensuring.pdf] Rothenberg, J. (2000) An experiment in using emulation to preserve digital publications. Den Haag: Koninklijke Bibliotheek [www.kb.nl/coop/nedlib/results/emulationpreservationreport.pdf] Rothenberg, J. (2003) Digital preservation summary. Presented at Practical Experiences in Digital Preservation Conference, National Archives, Kew, 24 April 2003 [www.nationalarchives.gov.uk/preservation/news/conference/media/rothenberg.pdf] Royal Statistical Society and UK Data Archive (2002) Preserving & sharing statistical material. Colchester: UK Data Archive. Rushbridge, A. (2004a) Recognising advances in digital preservation. DigiCULT.Info: A Newsletter on Digital Culture, 8, 3436 Rushbridge, A. (2004b) XENA: electronic normalising tool. DigiCULT.Info: A Newsletter on Digital Culture, 7, 3233 Russell, K. (1999) Why cant we preserve everything?: selection issues for the preservation of digital materials: debate and discussion at the Cedars Project Advisory Board Meeting, 28 May 1999, The British Library, St Pancras [www.leeds.ac.uk/cedars/ colman/ABS01.html] Saffady, W. (1993) Electronic document imaging systems: design, evaluation, and implementation. Westport, CT: Meckler Sanett, S. (2003) The cost to preserve electronic records in perpetuity: comparing costs across cost models and cost frameworks. RLG DigiNews, 7 (4) [www.rlg.org/legacy/ preserv/diginews/diginews74.html#feature2] Sassoon, J. (1998) Photographic meaning in the age of digital reproduction. LASIE, 29 (4), 515 Schoneld, R.C. et al. (2004) The nonsubscription side of periodicals: changes in library operations and costs between print and electronic formats. Washington, D.C.: Council on Library and Information Resources
234 Bibliography
Schller, D. (2000) Personal digital mass storage systems: a viable solution for small institutions and developing countries [www.unesco.org/webworld/points_of_views/ schuller.shtml] Schwirtlich, A. (2002) The functional approach to appraisal: the experience of the National Archives of Australia. Comma, 1/2, 5762. Searle, S. and Thompson, D. (2003) Preservation metadata: pragmatic rst steps at the National Library of New Zealand. D-Lib Magazine, 9 (4) [www.dlib.org/dlib/april03/ thompson/04thompson.html] Shenton, H. (2000) From talking to doing: digital preservation at the British Library. New Review of Academic Librarianship, 6, 163177 (Signicance): a guide to assessing the signicance of cultural heritage objects and collections (2001) Canberra: Heritage Collections Council Simpson, D. (2004a) Contracting out for digital preservation services: information leaet and checklist. London: Digital Preservation Coalition Simpson, D. (2004b) (comp.) Directory of digital preservation repositories and services in the UK [www.dpconline.org/docs/guides/directory.pdf] Sitts, M.K. (2000) (ed.) Handbook for digital projects: a management tool for preservation and access. Andover, Mass.: Northeast Document Conservation Center [www.nedcc.org/ digital/dighome.htm] Slats, J. (2004) Practical experiences of the Digital Preservation Testbed: ofce formats. Paper presented at the ERPANET File Formats for Preservation seminar, Vienna, 1011 May 2004 [www.erpanet.org/events/2004/vienna/index.php#papers] Slattery, O. et al. (2004) Stability comparison of recordable optical discs: a study of error rates in harsh conditions. Journal of Research of the National Institute of Standards and Technology, 109 (5), 517524 Smith, A. (1999) The future of the past: preservation in American research libraries. Washington, D.C.: Council on Library and Information Resources Smith, A. (2002) The cost of providing access. CLIR Issues, 27, 5 Smith, A. (2003) New-model scholarship: how will it survive? Washington, D.C.: Council on Library and Information Resources Smith, B. (2002) Preserving tomorrows memory: preserving digital content for future generations. Information Services & Use, 22, 133139 Smith, W. (1997) The National Library of Australias Pandora Project. Libri, 47, 169179 Smith, W. (2004) Wine on the web: Australian wine information on the web and its prospects for long-term preservation and access. Australian Academic & Research Libraries, 35 (2), 111128 Smith, W. (2005) Still lost in cyberspace?: preservation challenges of Australian internet resources. Australian Library Journal (forthcoming) Sochats, K. and Williams, J. (1992) Networking and communications desk reference. Carmel, Indiana: SAMS Society of American Archivists (1997) The preservation of digitized reproductions [www.archivists.org/statements/preservation-digirepros.asp] Soete, G.J. (1997) Preserving digital information. SPEC Kit 228. Washington, D.C.: Association of Research Libraries Steenbakkers, J.F. (2002) Preserving electronic publications. Information Services & Use, 22, 8996 Steenbakkers, J.F. (2004a) Digital archiving: a necessary evil or new opportunity? Serials Review, 30 (1), 2932 Steenbakkers, J.F. (2004b) Treasuring the digital records of science: archiving e-journals at the Koninklijke Bibliotheek. RLG DigiNews, 8 (2) [www.rlg.org/en/page.php? Page_ID=17068#article0] Strang, T. (2003) Choices and decisions. In Symposium 2003: Preservation of Electronic Records: New Knowledge and Decision-making, Ottawa, 1518 September 2003: preprints
Bibliography
235
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Stuckey, S. (2003) How to value a priceless asset: the experience of the National Archives of Australia or, What price archives and archivists? In GLAM: galleries, libraries, archives and museums: conference proceedings, Australian Society of Archivists 2003 National Conference, 1720 September 2003, Adelaide, pp.114120. OConnor, A.C.T.: Australian Society of Archivists Sturges, P. (1990) Access to information from the past: the birth of the amnesiac society? In Conference Proceedings: Papers Presented at the Australian Library and Information Association 1st Biennial Conference, Perth, Western Australia, September 30-October 5 1990, v.2, pp.295306. Canberra, ACT: ALIA Sullivan, J. (2001) True history of a writer. The Age, 6 October, Extra, 3 Sullivan, S. et al. (2004) Bringing hidden treasures to light: illuminating DSpace. Paper presented at VALA 2004 12th Biennial Conference and Exhibition, Melbourne, 35 February, 2004 [www.vala.org.au/vala2004/2004pdfs/17SHGY.PDF] Sullivan, S.J. (2004) ISO Standard development overview of the draft PDF/A. Paper presented at the ERPANET File Formats for Preservation seminar, Vienna, 1011 May 2004 [www.erpanet.org/events/2004/vienna/index.php#papers] Task Force on Archiving of Digital Information (1996) Preserving digital information. Washington, D.C.: Commission on Preservation and Access Task Force on the Artifact in Library Collections (2001) The evidence in hand. Washington, D.C.: Council on Library and Information Resources Taylor, G. (1996) Cultural selection. New York: Basic Books Thibodeau, K. (1999) Resolving the inherent tensions in digital preservation, Paper Presented at NSF Workshop on Data Archiving and Information Preservation, Washington, DC, 5 March 1999 Thibodeau, K. (2000) Certifying authenticity of electronic records, version 1, 19 April 2000. Unpublished Thibodeau, K. (2002) Overview of technological approaches to digital preservation and challenges in coming years. In The state of digital preservation: an international perspective, pp.431. Washington, D.C.: Council on Library and Information Resources Thibodeau, K., Moore, R. and Baru, C. (1999) Persistent object preservation: advanced computing infrastructure for digital preservation. Paper presented at DLM Forum99, Brussels, 1819 October 1999 [europa.eu.int/ISPO/dlm/fulltext/full_thib_en.htm] Thomas, D. (2004) Digital preservation at the National Archives [www.nationalarchives. gov.uk/preservation/digitalarchive/pdf/dpattna.pdf] Tibbo, H.R. (2003) On the nature and importance of archiving in the digital age. Advances in Computers, 57, 167 Townsend, S., Chappell, C. and Struijv, O. (1999) Digitising history: a guide to creating digital resources from historical documents. AHDS Guides to Good Practice [hds.essex. ac.uk/g2gp/digitising_history/index.asp] Tresize, P. (2002) Geoscience Australia: background report to the UNESCO Regional Consultation on the Preservation of Digital Heritage, Canberra, 46 November 2002. Tristram, C. (2002) Data extinction. Technology Review, October [www2.technologyreview.com/articles/02/10/tristram1002.asp] UNESCO (2003) Guidelines for the preservation of digital heritage, prepared by the National Library of Australia. Paris: UNESCO [unesdoc.unesco.org/images/0013/001300/ 130071e.pdf] UNESCO (2004) Charter on the preservation of the digital heritage. Paris: UNESCO [portal. unesco.org/ci/en/le_download.php/4cc126a2692a22c7c7dcc5ef2e2878c7Charter_en. pdf] University of Leeds. Representation and Rendering Project (2003) Survey and assessment of sources of information on le formats and software documentation [www.jisc.ac.uk/ uploaded_documents/FileFormatsreport.pdf] Upward, F. (2005) The records continuum. In Archives: recordkeeping in society, ed. S. McKemmish et al., pp.197222. Wagga Wagga: Centre for Information Studies
236 Bibliography
US-InterPARES Project (2002) InterPARES interpreted: a guide to ndings on the preservation of authentic electronic records [www.gseis.ucla.edu/us-interpares/pdf/ InterPARESInterpreted.pdf] Usai, P.C. (2000) Silent cinema: an introduction. London: British Film Institute Usai, P.C. (2001) The death of cinema: history, cultural memory and the digital dark age. London: British Film Institute Van Bogart, J. (1995) Magnetic tape storage and handling: a guide for libraries and archives. Washington, DC: Commission on Preservation and Access and National Media Laboratory. Also available at www.imation.com/government/nml/pdfs/AP_NMLdoc_ magtape_S_H.pdf Van de Velde, J. (1999) First Families 2001: creating and preserving Australian cultural heritage online. LASIE, 30 (2), 511 Van der Werf, T. (2002) Our digital heritage: how authentic should it be? Paper presented at VALA 2002: 11th Biennial Conference and Exhibition, Melbourne 68 February 2002 [www.vala.org.au/vala2002/2002pdf/21VndWrf.pdf] Van Diessen, R.J. and Van Rijnsoever, B.J. (2002) Managing media migration in a deposit system. IBM/KB Long-term Preservation Study Report Series, no. 5. Amsterdam: IBM; The Hague: Koninklijke Bibliotheek [www.kb.nl/hrd/dd/dd_onderzoek/reports/ 5-mediamigration.pdf] Van Wijngaarden, H. and Oltmans, E. (2004) Digital preservation and permanent access: the UVC for images [www.kb.nl/hrd/dd/dd_links_en_publicaties/publicaties/uvcist.pdf] Vogt-OConnor, D. (1999) Is the record of the 20th century at risk? CRM: Cultural Resource Management, 22, 2124 Vogt-OConnor, D. (2000) Selection of materials for scanning. In Handbook for digital projects: a management tool for preservation and access, ed. M. Sitts. Andover, Mass.: Northeast Document Conservation Center [www.nedcc.org/digital/dighome.htm] Walton, B. (2003) Accessibility and authenticity in digital records preservation. In Symposium 2003: Preservation of Electronic Records: New Knowledge and Decision-making, Ottawa, 1518 September 2003: preprints Warner, D. (2002) Why do we need to keep this in print? its on the Web . . . : a review of electronic archiving issues and problems. Progressive Librarian, 1920 [www.libr.org/ PL/1920_Warner.html] Waters, D.J. (1996) Archiving digital information. A presentation to the OCLC Research Library Directors Conference, Dublin, Ohio, March 12, 1996 Webb, C. (2000a) The likely impact(s) of digital technology on conservation/preservation in Australia. AICCM Newsletter, 75, 35 Webb, C. (2000b) The role of preservation and the library of the future. Paper presented at CONSAL 2000, Singapore, 2628 April 2000 [www.nla.gov.au/nla/staffpaper/ cwebb9.html] Webb, C. (2000c) Towards a preserved national collection of selected Australian Digital publications. New Review of Academic Librarianship, 6, 179191 Webb, C. (2002) Digital preservation a many-layered thing: experience at the National Library of Australia. In The state of digital preservation: an international perspective, pp.6577. Washington, D.C.: Council on Library and Information Resources Webb, C. (2004) The malleability of re: preserving digital information. In Managing preservation for libraries and archives: current practice and future developments, ed. J. Feather, pp.2752. Aldershot, Hants: Ashgate Webb, C. and Woodyard, D. (1998) Migration trials: from oppy disk to recordable compact disc (CD-R). LASIE, 29 (3), 2732 Weinberger, E. (1999) Digital objects as manuscripts: how to select material that is born digital for long-term preservation [www.cus.cam.ac.uk/~ew206/d-as-m-article] Weinberger, E. (2000) Towards collection management guidance. New Review of Academic Librarianship, 6, 6571
Bibliography
237
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Wheatley, P. (2001) Migration: a CAMiLEON discussion paper. Ariadne, 29 [www. ariadne.ac.uk/issue29/camileon/intro.html] Wiggins, R. (2001) Digital preservation: paradox & promise. Netconnect, Spring, 1215. [www.libraryjournal.com/article/CA106209] Wilson, A. (2003) Why the Dublin Core Metadata Initiative (DCMI) is important. DigiCULT.Info: A Newsletter on Digital Culture, 6, 3234 Woodyard, D. (1998) Farewell my oppy: a strategy for migration of digital information [www.nla.gov.au/nla/staffpaper/valadw.html] Woodyard, D. (1999) Practical advice for preserving publications on disk. Paper presented at Information Online and On Disc, Sydney, 21 January 1999 [www.nla.gov.au/nla/staffpaper/woodyard2.html] Woodyard, D. (2000) Digital preservation: the Australian experience. Paper presented at Positioning the Fountain of Knowledge: 3rd Digital Library Conference, Kuching 24 October 2000 [www.nla.gov.au/nla/staffpaper/dw001004.html] Woodyard, D. (2001) Data recovery and providing access to digital manuscripts. Paper presented at Information Online 2001, Sydney, 1618 January 2001 [www.nla.gov.au/ nla/staffpaper/2001/woodyard3.html] Woodyard, D. and Webb, C. (1997) Physical format electronic publications in the National Library of Australia: report on a preservation survey [www.nla.gov.au/nla/ staffpaper/cwebb6] Workshop on Research Challenges in Digital Archiving and Long-term Preservation (2003) Its about time: research challenges in digital archiving and long-term preservation: nal report. Washington, D.C.: National Science Foundation and Library of Congress Wyatt, A.L. (1990) Computer professionals dictionary. Berkeley: Osborne McGraw-Hill Yorke, S. (1997) Management of petroleum data records in the custody of Australian Archives. Archives & Manuscripts, 25 (1), 6273 Zwaneveld, E.H. (2000) Standards and new technology strategies to preserve content on magnetic and disc media. SMPTE Journal, 109, 628635
Index
Academic Research in the Netherlands Online 179 acceptable loss 7677, 80 access copies 201202 access, changing requirements 8 denition of 13 maintaining 184 accountability 18, 20 accrual accounting 189 action, need for 184, 192 active strategies 100 Agency to Researcher Digital Preservation project 211 AGLS metadata scheme 210 AHDS 173174 Alexa Internet 162 alteration for preservation 7576, 7980 AMIA 203 amnesiac society 25 analogue backups 102, 110111, 113, 119, 123125, 201 ANSI/NISO Z39.84 88 appraisal 53, 54, 5961 appraisal practice, applicability to digital materials 5960 archaeological data 2930 Archaeology Data Service 2930, 145, 173 archival le formats, developing 145146 archival practice 80 archival quality, denition 4 Archival Resource Keys 88 archival value 59
Archives Association of Ontario 60 Archives New Zealand 170 archiving, IT denitions 11 commercial services 156 archivists, and selection 5961 Australian, attitudes 79 preservation responsibility 20, 53, 54 ARKs see Archival Resource Keys artifacts, as information carriers 2, 3, 4, 7 preservation 4 artifactual value see digital artifacts; exhibition value Arts and Humanities Data Service see AHDS Association of American Publishers 88 Association of Moving Image Archivists see AMIA Association of Research Libraries 196 associational value, and selection 58 attributes, to be preserved 7598, 183, see also essential elements audiovisual archiving, case study 199203 selection 70 technology preservation 127 Australian Defence Force Academy xvi Australian Department of Family and Community Services 215219 Australian Digital Theses 179 Australian government archives 170 Australian Government Locator Service see AGLS
Index 239
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
Australian Institute of Aboriginal and Torres Strait Islander Studies 204 Australian National University 177 Australian practice xiv Australian preservation specialists, on strategies 105107 Australian Publishers Association 24,| 204 Australian Standard for Records Management 210 Australian War Memorial 204 Austrian On-line Archive 179 Authenticated Electronic Editions xvi, 93 authenticity 8891 denition 13, 15, 88 ensuring 37, 4950, 77, 81, 9091, 108, 184 loss of 27 pre-digital 8990 research 9596 threats to 90 awareness, of preservation 3132, 184 backup and restore see bit-stream copying backwards compatibility 110, 113, 152 BBC Domesday Project data 27, 29, 31, 130, 132133, 169 benign neglect 2, 10, 120, 182 Besser, H., on challenges 4849 bit-stream copying 102, 137 preservation 11, 79 born-digital materials 14 British Library 155, 191 research into digital preservation 193 brittle books 27 BS4783 5 Burrows, T., on selection 5657 business planning methodologies 187 business records, loss of 25 CAMiLEON Project 130, 132, 149, 153, 168169, 193 Canadian government records 29 carriers see digital artifacts CASL 181 CAUL 192 CD-R 122, 200203 CDs 202 storage and handling 121 CEDARS 178, 193 and costing data 188 and selection criteria 68 metadata scheme 85, 178
challenges xixii, 35, 37, 182193 compared to paper-based materials 37 choice see selection CICNet Journal Archive 30 Cinemedia 24 CLIR 219, 220 Cloonan, M., on paradigm shift 78 collaboration 157159, 183, 204, 211, 212, 213, 219220 requirement for 22, 32, 182, 203, see also international collaboration collection-based models 1 collections, declining importance 1, 36 combining strategies and practices 102, 104, 113, 114, 156, 182 commercial indifference 24, 51 community requirements, and authenticity 93 and essential elements 93 and selection 58, 65, 66 community will 105 compact disks see CDs compromise of data see data loss Computer History Museum 128 Computer Output Microlm 125 conceptual objects, denition 15 conservation, denition 4, 7 context preservation 4951, 6465, 8081, 90 continuous maintenance, need for 6, 37 continuum model xv and selection 6667 Conway, P., on paradigm shift 8, 5354 on selection 5354 Cook, T. 209 cooperation, see also collaboration; international collaboration lack of between librarians and archivists xiii copying 5, 123 for access 201202 copyright, law 63 of software code 132 Cornell University Library 191 corporate records see business records costing of preservation 184, 188, see also resources for preservation Council of Australian State Libraries see CASL Council of Australian University Librarians see CAUL Council on Library and Information Resources see CLIR Cox, R., on paradigm shift 5
240 Index
creators of information, preservation role 2324, 183 CSIRO xvi, 213 cultures, survival of 18 curation, denition 12 DARCA see Digital Archives and Records Coalition of Australia DAT tapes 200203 data, archiving see archiving creation 183 extent of loss 2629 extraction and structuring 102, 113 loss 2530 recovery 25, 3031, 113, see also digital archaeology recovery companies 30 data-centric approach 210 data-driven science 19, 23 dataVault 156 denitions, and professional associations 9 differences among information professionals 11 need for new 2, 47 pre-digital 4, 7, 10 revised 1216 revision of 1011 DELOS Programme 195 democracy, support of 18 dependence on playback machinery 37, 4445 deterioration, causes 36-51 of traditional library materials 25 DIAS 171, 172 DigiCULT 196 digital archaeology 102, 110, 112, 119, 126 Digital Archive 172 Digital Archives and Records Coalition of Australia 211 digital artifacts 3945 denition 39 handling of 40 lack of artifactual value 41 manufacturing quality 40 digital continuity see digital preservation Digital Curation Centre 12, 172173, 193 digital heritage, denition 13 Digital Image Archive of Medieval Music 179 digital information, volume of 56, 25, 36, 37 Digital Information and Archiving System see DIAS
digital materials, denition 13, 14 dependence on playback machinery 37, 4445 deterioration causes 3651, 40 fragility 37 life-span 36, 3940 selection for preservation 5373 Digital Object Identier 88 digital objects 4649, 77, see also digital materials denition 46 digital preservation, as a specialism 67, 189190 denition 13 future 196197 threats to 38 Digital Preservation Coalition 129, 174175, 191, 195, 219220 Digital Preservation Specialist, State Library of Victoria 67 Digital Preservation Testbed 143, 147, 176177, 193 Digital Recordkeeping Initiative 170 digital resources see digital materials Digitala Vetenskapliga Arkivet see DiVA digitization, products 14 selection for 67 DIRKS methodology 210 disaster preparedness 122, 126 dissertations see theses distributed custody see non-custodial models DiVA 147 DMR Consulting 214 do nothing strategy 102, 112, 119, 120 document persistence on the web 31 documentation, maintaining essential 4951 DOI see Digital Object Identier Dollar, C. 209 DSpace 164165 Duranti, L. 209 DVDs, storage and handling 121 Eastwood, T. 209 economic rationale 18, 188189 e-Depot 171172 education and training 189, 190191, 219220 e-journals see electronic journals electronic journals 30, 138, 158, 164 Electronic Literature Organization 31
Index 241
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
electronic publications, deposit system 169, 171172 electronic records 5, 208212, 212215 Electronic Resource Preservation and Access Network see ERPANET Elsevier Science 2324 e-mail 23 loss of rst 28 essential elements 70, 76, 9495 emulation 78, 102, 104, 110, 112, 113, 129133, 169, 177, 210 and encapsulation 154 application of techniques 29 encapsulation 102, 110, 113, 136, 153154, 212215 environmental conditions 120121 Epublish software 215217 equipment obsolescence see technological obsolescence e-repositories 23, 158, 164165 Ernst and Young Consulting 213 ERPANET 169170, 179, 193, 220 case studies 103104, 159, 177 technology options 108109 essential elements 76, 183, see also signicant properties denition 15, 88 identifying 94, 105 European Commission 169 European Union see DELOS Programme; DigiCULT European Visual Archive 179 evidential value, and selection 58 executable le formats 80 exhibition value, and selection 58 expanded range of stakeholders 103 expertise see skills eXtensible HTML see XHMTL eXtensible Markup Language see XML extent of preservation problem 2425 le formats 139140 proliferation 36, 140, 139141 registries 141143 restricting range of 144145 standardization 11, 102, 104, 111, 113, 114, 136, 143144, 148 sustainability criteria 143 lm archiving see audiovisual archiving formalization 112 formats see le formats fragility, and selection 58 Fujitsu 214 functional requirements, of systems 82
Functional Requirements for Evidence in Recordkeeping project 2526, 50, 9596, 209 funding see resources future, of digital preservation 181197 geoscience, preservation expertise in 31 Geoscience Australia xvi, 209 geospatial data 1 Global Digital Format Registry 142143 government records 29, 170, 172, 208212, 212215 growth rate of digital materials 36, 37 Handle System 88 handling media 122 HATII 170, 191 HD Rosetta 124 Hedstrom, M., on paradigm shift 5 historical memory 19 History Data Service 173 HTML 216218 HTTrack 207 Humanities Advanced Technology and Information Institute see HATII hybrid libraries 1 IASA 203 IBM 133, 169, 171 identity, denition 89 IIPC see International Internet Preservation Consortium individuals, preservation responsibility 2122 information, production and distribution of 3 information delivery practice 36 information management practice 36 information packages, denition 13, 1415, 82 information professionals, preservation denitions 9, 3132 information professionals, readiness for digital preservation 197 information society 23 initiatives 159161 institutional imperatives 1921 integrity, denition 8, 88 ensuring 37, 4950, 77, 90, 113, 184 loss of 27 threats to 4950 intellectual property rights 36, 37, 6264, 65, 68, 183
242 Index
International Association of Sound and Audiovisual Archives see IASA international collaboration 108, 110, 161169 International Internet Preservation Consortium 162, 163 International Organization for Standardization 146 internet, growth 23 Internet Archive 66, 72, 161163, 203, 206 InterPARES 50, 9697 and selection criteria 70 interpretability, maintaining 49 ISO 14721:2003 81 ISO 19005-1 146 ISTBAL 170 Jenkinson, Sir H. 20 JISC 178179, 191 JITM see Just In Time Markup Joint Information Systems Committee see JISC Jones and Beagrie, typology of strategies 110 journals, digitised 163164 JSTOR 158, 163164 Just In Time Markup 93 Kahle, Brewster 161 keep everything approach, sustainability 67, see also whole domain harvesting Kenney, A., on scalability 191 knowledge economy 18 knowledge, preservation of 18 Koninklijke Bibliotheek 2324, 130131, 133, 169, 171172 Landsat data 30 Lavoie, B., on economic issues 188189 legal deposit legislation 3, 57, 6364, 205 legal reasons for preservation 2021 librarians, preservation responsibility 1920, 53, 54 library materials, changing cultural value 5859 deterioration 25 Library of Congress 175, 194195 library practice, traditional selection criteria 5759 life-span of digital materials 36, 3940 LOCKSS Project 114, 138139, 158, 165166
logical objects see also digital objects denition 15 longevity, changing denition 8 long-term, denition 1112, 1516, 81 loss of authenticity 27 loss of integrity 27 loss of digital information 2526, 36 Lots Of Copies Keep Stuff Safe see LOCKSS Project LTO Ultrium tapes 201 magnetic media 4142 care and handling 42 life-span 42 manufacturers motivation 42 magnetic tape, storage and handling 121 mainstream, integration into 182184, 185187 maintenance of playback equipment 40 manufacturing quality 40 manuscript repositories, survey of 103 market forces 24, 51 market value, and selection 58 mass storage systems 102, 154156 Massachusetts Institute of Technology see MIT materials conservation, denition 4 MAVIS database 200, 201 media, acclimatization 121 denition 4 durability, ensuring 119, 122123 quality 122123 storage 120121 medium-term, denition 16 metadata 82, 8387, 102, 113, 214 extraction tools 86, 87 schemes 8487, 210 microlm, copying to 102, 123, 124, 125 Microsoft 197 migration 102, 104, 110111, 112, 113, 114, 136, 147153, 177, 201, 215219 need for 78 on request 153, 169 misinterpretation of evidence 75 MIT 164165 Monash University xv monitoring storage conditions 122 MP3 les 202 museums of computing 128 My File Formats 143 MyLifeBits 197
Index 243
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
NASA data 27 Nationaal Archief van Nederland 143, 170, 177 National Archives (UK) 142, 172 National Archives of Australia xv, 186187, 189 metadata scheme 86 strategies 208212 using XML 147 National Bibliographic Database 207 national collaboration 171177 National Digital Information Infrastructure and Preservation Program see NDIIPP National Film and Sound Archive 155, 199203 and PANDORA 204 national libraries, and legal deposit 57 cooperation with publishers 2324 National Library of Australia xivxv, 24, 187, 219220 and migration 150151 and sound preservation 40, 203 and web archiving 203208 Digital Object Storage System 155 guidance for small libraries 192193 international cooperation 159, 168 metadata scheme 8485 policy xixii, 37 preservation responsibility 21 research into digital preservation 193 survey 139140 technology listing 129 National Library of New Zealand, metadata extraction tool 86, 87 metadata scheme 8586 National Library of Norway 155 National Science Foundation (U.S.) see NSF National Space Science Data Center 30 NDIIPP 175176, 188 NEDLIB 85, 130131, 169 Networked European Deposit Library see NEDLIB Newham Museum Archaeological Service 30 nitrate lm deterioration 27 non-custodial models xv, 1, 5 non-solutions 118126 normalization 102, 103, 104, 211212 Northern Territory Library and Information Service, and PANDORA 204 NSF 194195
OAIS reference model 8182, 154, 167, 178, 182 obsolescence see technological obsolescence obstacles 181182 OceanStore 166 OCLC 88, 168, 188 Online Computer Library Center see OCLC Ontario Hydro nuclear power plant 29 Open Access Text Archive 162163 Open Archival Information System see OAIS operational mainstreaming 182184, 185187 optical disks 4344 care and handling 44 life-span 44 manufacturers motivation 44 Oxford Text Archive 173 Pacic and Regional Archive for Digital Sources in Endangered Cultures see PARADISEC Pacic region languages 177 PADI 159, 168, 179, 219220 and Safekeeping Project 220 PANDAS software 174, 207 PANDORA xv, 24, 72, 203208 paper, deterioration 25 printing to 102, 123, 124, 125 paradigm shift 710, 41, 102 PARADISEC 177 passive strategies 100 PDF, as standard 104, 214, 216 PDF/A 145146 performance 92 Performing Arts Data Service 173 persistent identiers 8788, 110 persistent object preservation 102, 111 Persistent Uniform Resource Locator 8788 personal correspondence 23 personnel required 67 physical objects see also digital artifacts; media as information carriers see artifacts, as information carriers denition 15 Pilgrim Trust Preservation Awards 172, 175 Piltdown man 75
244 Index
playback equipment, dependence on 37, 4445 maintenance 40 Pockley, S. 24 policy development 102, 119, 125, 183 Portable Document Format see PDF practices, criteria for effective 107110 denition 100 for preservation 116 PREMIS Working Group 85 survey 182 pre-print archives 23 preservable essence see essential elements; signicant properties preservation, awareness of issues 3132, 184 conceptualization of 89 denitions 1013 extent of problem 2425 paradigms 116 pre-digital paradigm 2, 34, 5 program, denition of 13, 16 purposes 1733 resources, justifying requests for 26 27 responsibility for 1733, 37, 158, 182 preserve objects approaches 136156 preserve technology approaches see technology preservation Preserving Access to Digital Information see PADI Preserving and Accessing Networked Documentary Resources of Australia see PANDORA principles, denition 99 for preservation 115 printing to paper 102, 123, 124, 125 privatization, of preservation 21 professional, associations, preservation denitions 9 professional imperatives 1921 professional practice, Australian xiv project-based management 185 PRONOM 142 proprietary le formats 80, 141 provenance 90 Public Record Ofce of Victoria 147, 211, 212215 publishers, preservation responsibility 2324 PURL see Persistent Uniform Resource Locator Quadriga workstations 200
recordkeeping 18 practice 53, 54, 5556, see also archivists Recordkeeping Metadata Project xv redundancy 102, 114, 166, 182, see also replication reformatting, and paradigm shift 7 refreshing 5, 104, 137138 denition 137 regional collaboration 169170 relative humidity, in storage areas 120121 replication 138139 denition 137 research data loss, Canada 30 Research Libraries Group 167168 survey 3738, 103, 107, 140 research, into digital preservation 193196 quality, measurement of 26 resources for preservation 2627, 37, 105, 158, 184, 185, 187189 responsibility for preservation 1733, 37, 158, 182 restoration, denition 4 risk management 137, 187 RLG see Research Libraries Group RMIT University 24 Rothenberg, J. 39 and emulation 131 on migration 151 on strategies 107, 112 RTF 216217 SAA see Society of American Archivists Safekeeping Project 220 scalability, of solutions 185, 191193 scholars, preservation awareness 32 preservation expectations 23 Schweizerisches Bundesarchiv 170 scientic data 19, 23, 30 ScreenSound Australia see National Film and Sound Archive sectoral collaboration 177179 seismic data 208209 selection, changing denition 8 criteria 5762 decision-making framework 6972 decisions, bases for 5657 debate about 5557 for digitization 67 for preservation 5373, 113, 182, 183 importance of 5556 in archives 5961 in libraries 5759, 205206
Index 245
1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111
inuence of preservation cost 73 vs. keep everything approach 72 separation, of content from technologies 103 short-term, denition 16 signicance, changing concept 7273 signicant properties 68, 88, 9195, see also essential elements skills, development of 184185, 189191 required 190 small institutions 191193 small-scale projects 215219 social science data 1,150 social security law, access to 215219 societal benets 1819 Society of American Archivists, preservation denitions 9 software, dependence on 4849, 7778 obsolescence see technological obsolescence repositories 102 sound archives 177 sound archiving see audiovisual archiving sound recordings, transfer from analogue to digital storage 40 standards, and standardization 81, 86, 102, 105106, 110, 113114, 119, 125126, 182183, see also le formats, standardization; normalization change 4445 for sound archiving 200 State Library of New South Wales 127 and PANDORA 204 State Library of South Australia, and PANDORA 204 image digitisation 4445 State Library of Tasmania, and PANDORA 204 State Library of Victoria, and PANDORA 204 survey of digital materials 129 State Library of Western Australia, and PANDORA 204 State Records Authority of New South Wales 210, 211 storage, and handling practices 119, 120122 conditions 120121, 122 Storagetek systems 200 strategies 99116, 212 appropriateness 108 criteria for effective 107110 denition 99 feasibility 108
practicality 108 sustainability 105, 108 Sudwestrundfunk 155 surveys, of digital holdings 129 of international practice 104 of National Library of Australia holdings 139 of RLG digital holdings 3738, 103 of scientic publishers 104 of US manuscript repositories 103 sustainability, of strategies 105, 108, 114 systems design 1011 Task Force on Archiving of Digital Information 22 technological obsolescence 3651, 4445, 4648, 78, 111, 127 monitoring 128129 technology museums see technology preservation technology preservation 78, 102, 104, 110111, 113, 117, 126134 technology watch 102, 110, 127, 128129 temperature, for storage areas 120121 Testbed Digitale Bewaring see Digital Preservation Testbed theses, selection criteria 72 Thibodeau, K., on strategies 108 typology of strategies 111112 Tibbo, H., on future of digital preservation 197 TIFF 214 training see education and training transferring to microlm 102, 123, 124, 125 trusted digital repositories 9798 trustworthiness, certication 98 of records 81 typologies, of initiatives 159161 of strategies 100, 110116 UNESCO 166167, 192 typology of strategies 112114 Uniform Resource Name 87 United States 1960 Census data 28 United States government records 29 Universal Preservation Format 146 Universal Virtual Computer 102, 104, 113, 133134 University of British Columbia 209 University of Leeds 169 Representation and Rendering Project 141142 University of Melbourne 177, 213 University of Michigan 169
246 Index
University of New England 177 University of Pittsburgh 2526, 50, 9596, 209 University of Sydney 177 URN see Uniform Resource Name user need, and selection 5758, 59 UVC see Universal Virtual Computer variety of digital materials 6 VEO 214 VERS xvxvi, 212215 and encapsulation 154 and XML 147 metadata specications 214, 217 VERS encapsulated object see VEO version migration 113, 148, 151152 Victorian Electronic Records Service see VERS videotapes, selection criteria 71 Vietnam War combat area casualty data 27 viewers 102, 113, 114, 153 Viking Mars mission data 27 Visual Arts Data Service 173
Walton, B., typology of strategies 110111 Wayback Machine 162 web archiving 161163, 168, 174, 203208, 212 selective approach 203, 205207 Web Archiving Consortium 174 web domain harvesting 02, 203, 205 web persistence 31 web sites, essential elements 76, 95 selection criteria 71, 205206 WebZIP 207 Wheatley, P., on migration 148149 whole domain harvesting 72, 203, 205 World Trade Centre, bombing, 1993 25 Wotsits Format 143 Xena 211212 XHTML 218 XML 146147, 177, 211212, 213214, 216218 XML Electronic Normalising of Archives see Xena Yale University Library 150

Preserving Digital Materials

Uploaded by

Copyright:

Available Formats

Preserving Digital Materials

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Preserving Digital Materials

Uploaded by

Copyright:

Available Formats

Preserving Digital Materials

Preserving Digital Materials

Preserving Digital Materials

K G Saur Mnchen 2005

1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111

1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111

1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111

201 202 203 203 208 212 215 219

1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111

1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111

1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111

1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111

What is Preservation in the Digital Age? Changing Preservation Paradigms

2 What is Preservation in the Digital Age? Changing Preservation Paradigms

The need for a new preservation paradigm 3

1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111

Source: Internet Software Consortium (www.isc.org)

Figure 1.1 Growth Rate of the Internet

The need for a new preservation paradigm

4 What is Preservation in the Digital Age? Changing Preservation Paradigms

The need for a new preservation paradigm 5

1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111

6 What is Preservation in the Digital Age? Changing Preservation Paradigms

1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111

8 What is Preservation in the Digital Age? Changing Preservation Paradigms

Preservation denitions in the digital world 9

1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111

Preservation denitions in the digital world

What is Preservation in the Digital Age? Changing Preservation Paradigms

Preservation denitions in the digital world 11

1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111

What is Preservation in the Digital Age? Changing Preservation Paradigms

1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111

What is Preservation in the Digital Age? Changing Preservation Paradigms

What exactly are we trying to preserve?

Preservation denitions in the digital world 15

1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111

How long are we preserving them for?

What is Preservation in the Digital Age? Changing Preservation Paradigms

What strategies and actions do we need to apply when we preserve them?

1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111

Why do we Preserve? Who Should do it?

Why do we Preserve? Who Should do it?

Why preserve digital materials?

1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111

Why do we Preserve? Who Should do it?

1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111

Why do we Preserve? Who Should do it?

1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111

Why do we Preserve? Who Should do it?

The extent of the preservation problem

The extent of the preservation problem 25

1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111

Why do we Preserve? Who Should do it?

How much data have we lost?

How much data have we lost? 27

1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111

Why do we Preserve? Who Should do it?

How much data have we lost? 29

1111 2 3 4 51 6 7 8 9 10 1 2 3111 4 5 6 7 8 9 20111 1 211 3 4 5 6 71 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 44 45 46 47 48 49 50111