On spurious complexity

Thimbleby

Computing Science Middlesex University, UK

Computing
 complex
 unreliable
 we depend on them
 buy upgrades to go even faster

Tamagochi
 fashion accessories
 appear very simple
 have a life of their own
 purpose: difficult to use
 badly engineered (need reset buttons)
 yet, people buy them eagerly...
 mass produced, with little intrinsic value
 not difficult to design

Computing provides
 autonomy
 creativity
 control
 performing chores
but:
 behaviour is difficult to identify & comprehend
 unlikely achieve intended goals
 support and are promoted by a social structure

Programming
 linear
 grow sub/linearly with the programmer typing (some typing may be deletions)
 number of things a program can do grows exponentially with the number of
interactions
 the system state is often unknown, the outcome is therefore unforseeable
 Humans
 have a bounded raionality: finite computation speed and finite lifetime (upper bound
on cognitive resources)
 skilled people build programs that sometimes may not understand
 understanding a system overloads our cognitive resources
 it is therefore difficult to evaluate it

Media equation
Reeves e Nass 1996
 we take media as a reality (evolutionary speaking)
 everything we perceive is taken as real
 a real program behaving like one demonstrated would work everywhere in the same
way...yet the demonstration has shown a single trace
 we cannot distinguish between a simulation and the real thing

Theory and theatre

 common greek roots (objects of study/objects to study)
 media equation: willing suspension of our critical faculties
 in computing: it is hard to watch a demonstration and ask about off/stage questions
 we believe the trace was typical of the general behaviour of the program

Attitudes
 collusion (we collude with bad design)
 successful dreams (optimistic about scenarios)
 lottery effect (few successes imply many positive views on computer usage)
 review culture (technical reviews are inaccurate, do not fully test, list new features)
 succesful children ...dumb adults (playing vs rationalizing)
 realism-reality gap (it looks realistic, i.e. it is real)
 Oracle effect (experts underestimate complexity)
 one way to use computers ("experts" show the way...only 1 way)
 automation is not computation (speeding up procedures is not necessarily a good
thing)
 Lyability Problem: quando si apre il sigillo le aziende non si assumono alcuna
responsabilità nell'abilità dell'utente se esso sia in grado o meno di utilizzare con le
proprie conoscenze il programma).
 inertia (liability problem, there is no reason to improve since we all benefit by not
improving)
 usability is the user's problem (quote R. Nader's book Unsafe at any Speed) people
are told to read handbooks so feeding the industry of training and consultancy
 Possible solutions
 good design + better handbooks
 construct the initial user manual
 find problems (technical authors do this)
 fix the design (the user manual is a direct indicator of design areas needing attention)
and repeat : each step improves the product

Manuals
 it is easy to write vague, inexact and misleading manuals
 manuals should be short, complete and sound
 there may be different users for a manual: a mathematician, an electronic engineer, a
programmer

Justice
 Aristotle defines justice as the act of giving a person good
 to do good design is to be engaged in an act of justice
 justice is the maintenance of rights(for the user)
 just design is to maintain those rights
 good manual writers contribute to a just world

Theory of justice
Rowls, 1972
 justice as fairness: a system of rules that would be designed by people under a "veil
of ignorance" of whether and to what extent those rules applied to themselves
 fairness applies to designers as well as users
 designers generally do not act justly

Design
 good design is engagement with justice
 be conscious of the way in which users will operate the product
 user concurrent engineering with user manuals
 put oneself into the many ¹ roles of usage
 justice is the only virtue that can be achieved by accident (Aristotle)
 integrity cannot be reached by accident
 justice is an outcome and holds if it fairly affects others
 optimistic view: market will help ensure good design
 pessimistic view: superior technical systems fail due to economic and social forces

Conclusions
 computing systems are complex
 since they are difficult to evaluate, we take them as objects for uncritical consumption
 it suits almost everybody: manufacturers make money, book publishers sell
dummies, marketing people have a lot to advertise...we all swallow
 concurrent and iterative manual design

Human Computer Interaction

 ACM: Interaction Journal
 SIGCHI: around 6000 members:annual conference draws 2500 attendees
 Italian Chapter - HCItaly (CNR, Feb. ‘99)
 Usability Prof. Ass.
 American Soc. for Information Science

Human Factors of Interactive

 present situation
 goals of system engineering
 goals of user-interface design
 accommodation of human diversity
 goals for our profession
 references

Present Situation
 computer pervasiveness implies many, different users
 which, in turn, means many, different interfaces
 web, games, databases, simulation, business, training, education, etc. are but a few
of the applications
 impacting on: info organization, query languages, 3d representations, animation,
direct manipulation, telepresence, virtual realities...(sw)
 novel keyboards, high resolution color displays, speech i/o, gestural inputs, force-
feedback devices, touchscreen & stylus...(hw)

Goals of System Engineering

 to serve the users
 to design effective systems
 in good design, the interface becomes invisible:
US Military Standard for Human Eng. Criteria

 allow for required performance by operator, control and maintenance personnel

 minimize skill and personnel requirements and training time
 achieve required reliability of personnel-equipment combinations
 foster design standardization within and among systems

Main issues in design

 proper functionality (task analysis is central)
 reliability, availability, security and data integrity (across commands)
 standardization, integration, consistency and portability (across applications,
platforms and op. systems)
 schedules and budgets

Goals of User-Interface Design

 If the previous issues have been addressed then developers must focus on design
and testing
 Careful determination of the user community and benchmarks for testing
 five measurable human factors:

 time to learn
 speed of performance
 rate of errors by users
 retention over time
 subjective satisfaction

Tradeoffs
 lengthy learning favours task-performance
 if low error rate then performance is hindered
 subjective satisfaction may be a key point
 design teams negotiate guidelines
 paper prototype/mock up prototype

Motivations for Human Factors Design

 Life critical systems
 Industrial and commercial uses
 Office, home and entertainment applications
 Exploratory, creative and cooperative systems

 CSCW, distance learning, videoconferencing...

 high motivation but also high expectation...
 the computer should vanish...
 Accommodation of Human Diversity
 Physical abilities and physical workplaces
 American National Standard for Human Factors Engineering of Visual Display
Terminal Workstations

 Work-surface and display support height

 Clearance under work surface for legs
 Work-surface width and depth
 Adjustability of heights and angles for chairs and work surfaces
 Posture-seating depth and angle, back-rest height and lumbar support
 Availability of armressm footrestsm and palmrests
 Use of chair casters

Room layout and sociology of human interaction

 multiple workstations - encourage/hinder social interaction
 privacy versus open space office
 physical design of workplaces belongs to ergonomics

Cognitive and perceptual abilities

 Ergonomics Abstracts:

 Short-term memory
 Long-term memory and learning
 Problem solving
 Decision making
 Attention and set (scope of concern)
 Search and scanning
 Time perception

Factors affecting perceptual and motor performance

 Arousal and vigilance
 Fatigue
 Perceptual (mental) load
 Knowledge of results
 Monotony and boredom
 Sensory deprivation
 Sleep deprivation
 Anxiety and fear
 Isolation
 Aging
 Drugs and alcohol
 Circadian rhythms

Personality Differences
 Men/women
 Interaction styles
 Pace of interaction
 Graphics versus Text
 Dense versus sparse
 Step-by-step versus all-at-once

Myers-Briggs Type Indicator

 Extroversion vs introversion
 Sensing versus intuition
 Perceptive versus judging
 Feeling versus thinking
Many other psychological scales have been suggested...
 Risk taking versus risk avoidance
 Internal versus external locus of control
 Reflective versus impulsive behaviour
 Convergent versus divergent thinking

Cultural and international diversity

 Cultural, ethnical, racial, linguistic background
 User interface design concerns for internationalization include:

 characters, numerals, special characters

 left to right vs right to left (vertical vs horizontal)
 date and time formats
 numeric and currency formats
 telephone numbers and addresses
 names and titles (Mr., Mrs., Mme., Dr.)
 Social security, national identification, passport number
 capitalization and punctuation

and...

 sorting sequences
 icons, buttons, colours
 pluralization, grammar, spelling
  etiquette, policies, tone, formality, metaphors

 Companies should run usability studies in each country, culture and language
community

Users with disabilities

 for low vision (even blind)
 for hearing impairments
 for mobility impairments
can all be accommodated using:
 special i/p devices
 speech recognition
 eye-gaze control
 head-mounted optical mouse

1. were pioneered for disabled users

Computer curbcuts
Cheaper to handle if planned (rather than added)
 learning-disabled children (2% of school population)
 proceduresm directions and verbal content at levels and in formats making them
accessible even to poor readers
 response requirements do not allow students to complete programs without engaging
with target concepts
 design feedback sequences that explain the reasons for sutdents’errors and that
lead students through the processes necessary for responding correctly
 incorporate reinforcement techniques that capitalize on students’ sophistication with
out-of-school materials

Elderly users
 growing population of senior citizens
 writing accounting, education, entertainment, social interaction, communication and
challenge
 increased access of the society
 increased participation through networks
 improved chances for productive employment
 government agencies, universities, medical centers, law firms could use senior
experience
 e-mail connection enriches family life
examples
 larger street signs
 brighter traffic lights
 better nighttime lighting
 larger fonts
 higher display contrast
 easier-to-use pointing devices
 louder audio tones
 simpler command languages
golden age software + kidware: large opportunities

Goals for Our Profession

 influence academic and industrial researchers
 provide tools, techniques and knowledge for commercial systems implementors
 raise the computer consciousness of the general public

Influencing academic and industrial researchers

 early methods were based on introspection and intuition

 lack of validity, generality, precision

 reductionism scientific method (psychologically oriented) performs the following

steps:

1. understanding of a practical problem and related theory

2. lucid statement of a testable hypothesis
3. manipulation of a small number of independent variables
4. measurement of specific dependent variables
5. careful selection and assignment of subjects
6. control for bias in subjects, procedures, and materials
7. application of statistical tests
8. interpretation of results, refinement of theory, and guidance for experimenters

Controlled Experiments
 have limitations:

 difficult or expensive to find adequate subjects

 right laboratory conditions
 extreme performances may be overlooked
 anecdotal evidence/individual insights may be ignored

 yet controlled experimentation is productive and may be adapted to each situation

 field or test experiments should be carried out
 individual insights of researchers, designers, participants should be documented

Human Interfaces
 need to understand and measure human performance
 awareness to human-factors issues
 knowledge and techniques of traditional psychology to be used for studying human-
computer interaction
 information science, business and management, education, sociology, anthropology
are all contributing and benefiting by their study of human-computer interaction

Directions of Study
 Reduced anxiety and fear of computer usage
 Graceful evolution

 methods to smooth the transition from novice to knowledgeable user to expert

 adaptation and evolution are open topics addressed in the re-design of control
panels

 Specification and implementation of interaction: building tools reduce 10% if the tool
is adapted to the task
 Specification and implementation of interaction: building tools for interface
construction
 Direct manipulation - which is the best metaphor? - new forms: visual languages,
spatial visualization, remote control telepresence virtual reality
 Input devices - high-precision touchscreen, stylus, pen, voice, eye-gaze, gestual;
mouse, dataglove, force feedback joystick
 Online assistance - limited understanding for novices, knowledgeable and expert
users - JIT: just-in-time training
 Information exploration - navigation, browsing, searching of multimedia digital
libraries - filter

Tools, techniques and knowledge for system

implementers
 new tools (environments & prog. languages)
 guidelines
 iterative usability studies
 feedback from users
 online user consultants

Computer consciousness of the general public

 media boosts computer power
 yet people feel uncomfortable about the computer
 people are afraid of mistaking
 people feel incompetent
 people feel threatened
 all these fears are generated by bad design
 designers should be more concerned about users

Practioner’s summary
 thorough task analysis
 attention to : reliability, availability, security, integrity, standardization, schedules &
budgets
 evaluation of design alternatives: short learning times, rapid task performance, low
error rates, ease of retention, high user satisfaction
 for refined design test: accomplishment of goals, expert reviews, usability tests,
acceptance tests

Theories, Principles, and Guidelines

 emerging designers follow some guidance
high level theories and models
middle-level principles
specific and practical guidelines
 user interface displays could be improved
clutter, complex and tedious procedures, inadequate functionality, inconsistent
sequences of actions, insufficient informative feedback

High Level Theories

 explanatory
observe behavior, conceive designs, compare high-level concepts of 2 designs,
describe activity
 predictive
compare proposed designs for execution time or error rates, motor-task
predictions (keystroking, pointing times), perceptual theories predict reading
times, predicting performance on cognitive task is hard (ratio to perform tasks
between novice and expert may be 100/1)

Taxonomies
 input devices
direct vs indirect, linea vs rotary
 tasks
  structured vs unstructured, controllable vs immutable personality styles

 convergent vs divergent, field dependent vs independent

 technical aptitudes

 spatial visualization, reasoning

 user experience levels

 novice, knowledgeable, expert

 user-interface styles

 menus, form filling, commands

Many theories...
 compete for attention
 continuous refinements
 extended by critics
 applied by eager and hopeful designers
 implementers must keep up with developments not only in s/w tools but also in
theories

Subject satisfaction
 researchers in media and advertising recognize the difficulty in predicting emotional
reactions
 theoretical predictions are combined with intuitive judgements + extensive market
testing
 broader theories: small group behavior,organizational dynamics, sociology of
knowledge
 barriers to new technology may be analyzed via social psychology and anthropology
 coming up with a god theory is not simple

Conceptual, semantic, syntactic and lexical model

Foley & Van Dam (1970)

 conceptual: user’s mental model of the interactive system

 a line editor models a word processor program

 semantic: describes meanings conveyed by the user’s command i/p + comp. display
 syntactic: how the units are assembled into a complete sentence
 lexical deals with device dependencies + precise mechanisms by which a user
specifies the syntax

Advantages
 the above model is clear since it is top-down
 matches the s/w architecture
 allows useful modularity during design
 designers move from conceptual to lexical and carefully record mappings between
these 2 levels

Keystroke-level model
 attempts to predict performance times for error-free task execution
 different activities are considered: key-stroking,pointing, homing, drawing, thinking,
waiting for system response
 the sum of all insured times provides performance times
 these models hold for experienced users and error-free performance
 no emphasis on learning, problem solving, error handling, subjective satisfaction and
retention

GOMS:goals, operators, methods & selection rules

Card, Moran and Newell (1980-3)

 users formulate goals (edit documents) and subgoals (insert word) by using methods
or procedures

 move cursor by following a sequence of arrow keys

 operators are elementary perceptual, motor or cognitive acts whose execution is

necessary to change any aspect of the user’s mental state or to affect the task
environment
press up-arrow key, move hand to mouse, recall file name...
 methods-selection rules these rules are used to choose one possible way of
achieving the task

Again on GOMS
Kieras and Polson (1985)
 Kieras and Polson formalized GOMS by using production rules
 via these rules predictions of learning and performance time were given when
interacting with a text editor through 5 different actions:

 insert, delete, copy, move and transpose

Natural GOMS
Kieras (1988)
 "the GOMS analysis did not explain how the notation works, it is clumsy, detached
from the underlying cognitive theory"
 GOMS was refined into Natural GOMS Language (NGOMSL)
 Find when the task-analyst must make: a judgement call

 assumptions about how the users view the system

 bypass a complex hard-to-analyze task
 check for consistency

Method descriptions
Elkerton and Palmiter (1991)

 They applied NGOMSL to implement on-line Help

 introduced method descriptions breaking down the actions necessary to accomplish
a goal into steps

 decide, accomplish, report goal accomplished

 introduced selection rules whenever alternative methods exist

 empirical evaluation with 28 subjects proved that NGOMSL version of help halved
the time to complete information searches

Stages of action
Norman (1988)

 7 stages of action to model HCI:

 forming the goal

 forming the intention
 specifying the action
 executing the action
 perceiving the system state
  interpreting the system state
 evaluating the outcome

 similar to Foley and van Dam’s separation of concerns:

 conceptual intention - reformulation into commands

 syntax construction - production of lexical tokens

Norman’s contribution
 taking care of

 cycles of action
 evaluation

 the process of action is considered dynamically (i.e. in evolution)

 alternative models only consider what is in the user’s mind at execution time
 Two new concepts:

 gulf of evolution which separates user’s intentions from allowable actions plus
 gulf of evaluation separating system’s representation from user’s expectation

Four principles
 Norman’s four principles for good design

 state and action alternatives should be visible

 there should be a good conceptual model with a consistent system image
 the interface should include good mappings that reveal the relationships between
stages
 the user should receive continuous feedback

 Study errors - they often occur when moving from goals to intentions to actions and
to executions

Exploring the interface

Polson and Lewis (1990)

 when users explore an interface and try to accomplish their goals, they pin-point 4
critical points where failures may occur:
  users can form an inadequate goal
 users might not find the correct interface object because of an incomprehensible
label or icon
 users may not know how to specify or execute a desired action
 users may receive inappropriate or misleading feedback

Franzke (1995)

 the bottom 3 failures may be prevented by improved design or time consuming

experience

Consistency through grammars

 consistency (coerenza!) is elusive with multiple levels
 it may be even positive to be inconsistent !
 a command language or set of actions should be:

 orderly - predictable - describable by a few rules

 easy to learn - easy to retain

 these overlapping concepts are shown by an example showing 2 kinds of

inconsistencies...

Consistencies-Inconsistencies
Consistent + Inconsistent A Inconsistent B
delete/insert character delete/insert character delete/insert character
delete/insert word remove/bring word remove/insert word
delte/insert line destroy/create line delete/insert line
delte/insert paragraph kill/birth paragraph delte/insert paragraph

 all actions are the same in + but vary in version A

 the inconsistent verbs are all acceptable but their variety suggests they will be more
difficult to learn, to remember, will slow down users, will be error-prone
 version B is more malicious: only one inconsistency (remove) but may be easier to
remember...

Task Action Grammar

Payne and Green (1986)
 expanding on Reisner (1981) they addressed:

 multiple levels of consistency (lexical, syntactic and semantic)

 aspects of completeness (complete set of tasks

 once the full set of task-action mappings is written down, the grammar of the
command language can be tested for completeness

Command Language Grammar

 e.g. a TAG definition of cursor control:

 move-cursor-one-character-forward
[Direction = forward, Unit=char]
 move-cursor-one-character-backward
[Direction = backward, Unit=char
 move-cursor-one-word-forward
[Direction =forward, Unit=word]
 move-cursor-one-word-backward
[Direction =backward, Unit=word]

High rule schemas describing the commands syntax

These schemas generate a consistent grammar

move cursor one character forward CTRL-C

move cursor one character backward ESC-C
move cursor one word forward CTRL-W
move cursor one character backward ESC-W

Again on consistency
 notation and approach (for TAGS) are flexible and extensible
 consistency is subtle, multiple levels, may also hinder some implementation details -
Reisner (1990)
 understanding consistency is instrumental for implentors, researchers and designers
- Grudin (1989)

Widget level theories

 reductionism approach may be fallacious, i.e. details may become misleading in the
evaluation of a GUI
 validity of simple summations of time periods may be questionable
 alternatively, one may use a model based on widgets (interface components)
 layout appropriateness (frequently used widgets should be adjacent, left-to-right
sequence should be matched to the task-sequence description,...)
 higher-level patterns appear by widget composition

Object-Action Interface Model

 syntactic-semantic model of human behaviour - Shneiderman (1980)
 this model describes programming, database manipulation facilities, direct
manipulation
 a distinction was made between

 acquired semantic concepts (delete - copy)

and

 rote-learned syntactic concepts (function keys)

 the first were stable in memory and well organized, the second were arbitrary and
had to be rehearsed to be maintained

Objects and Actions: differences

 task domain concepts (stock-market portfolios)
 computer domain concepts (folders)
OAI model
 GUIs have replaced command languages substituting complex syntax with direct
manipulation
 emphasis is now on the visual display of user tasks objects and actions

 stock-market portfolios could be represented by leather folders with engraved

certificates
 actions represented by trashcans, shelf icons,etc
 Object-Action design
 action syntax is easier than command language expressions even if precedence
rules must be known (file/folder to the trashcan and not viceversa)
 mouse clicking, mouse retention, gestures also obey rules
 design starts with a clear identification of the task/s - including the universe of real
world objects + user intentions + actions required
 high level task objects could be stock-market statistics, a photo library, a scientific
journal

Task and Interface concepts

OAI Model
 the two pictures show

 the objects of the universe (shelves, cupboards, books in a library)

 the actions satisfying the intentions of the user
 the objects of the interface visualizing (through pixels)
 the actions to be performed by the user (via mouse clicks)

 in this way, the interface is easy to learn and to use since it maps the world domain
with the metaphoric domain
 it focuses on task objects and actions and on interface objects and actions
 OAI reflects design at high level as when programmers use widgets in usr-interface-
building tools
 standard widgets have a simple syntax:click, double-click, drag, drop
 OAI follows the object-oriented approach
 Hierarchies
 when problems are complex: break them down!
 intentions may be decomposed into smaller action steps

 building: surveying, building the frame, raising the roof

 symphony: movements, measures, notes
 baseball: innings, outs, pitches

 people learn task objects & actions independently of their implementation

 people learn by studying & practicing

Application domains
 designers must learn via

 training courses
 books
 interviews

 designers generate a hierarchy of objects and actions to model the user’s tasks
 the model is the basis for designing the interface objects and actions + their
representation in pixels on the screen, in physical devices or audio cues
 users must firstly become proficient in their task domain
 next, they may learn the equivalent computer program

Interface objects
 the interface includes hierarchies of objects and actions at high and low levels
 storage is a high level concept: computers store information

 by means of the directory and files (objects)

 a directory is made of entries (lower level objects)
 each entry is made of a name, length, creation date, owner, access control,...
 each file has lines, fields, characters, fonts, pointers, binary numbers,...

Interface actions
both high and low level actions like
creating a text data file
load, insertion, save actions
storing a file, backup on one of many disks, applying access control
rights,...
permissible file types, sizes, error conditions, responses to h/w or s/w errors
and finally...clicking on a pull-down menu

Familiar examples
 a designer will build interface objects and actions based on familiar examples
 tune those objects and actions to fit the task
 for a real estate business, geographical maps and houses will be available as well as
their properties, cost, distance, size and location (familiar concepts) will be mapped on
the screen
 to explain "saving a file", icons representing a disk drive and the directory will show
where the file will be stored
 a demonstration will be performed to enable the user a logical understanding of the
process

Metaphors
 they map a meaning from one known domain to another one
 they may be abstract, concrete or analogical
 they are used to avoid long, tiresome training of new concepts
 most icons user representations which are visual metaphors (which, in itself, is a
metaphor)
 interface objects and actions have a logical structure which is easy to memorize in a
stable way

Bottom-up modelling
 task objects made explicit
 user’s task actions laid out clearly
 next, interface objects are identified
 and interface actions follow...with the OAI model
 many years ago users had to remember device dependent details (format instruction
in Fortran, number of i/o device to be deployed,etc.)
 or, which action deletes a character: delete, backspace, CTRL-H, CTRL-G, CTRL-D,
rightmost mouse button or Escape
 which action inserts a new line after the third text line: CTRL-I, INSERT KEY, I3, I 3,
3I,...

Remembering...
 problem 1: details vary across computer platforms
 problem 2: arbitrariness of minor design features reduces the effectiveness of paired-
associate learning
 repeated rehearsals for rote memorization
 moreover, syntactic knowledge is hampered by the lack of a hierarchical or modular
structure to cope with complexity

Example
 within e-mail:

 press RETURN to terminate a paragraph

 CTRL-D to terminate a letter
 Q to quit the e-mail subsystem
 logout to terminate the session

 for the novice these similar termination commands bear no logical connection

Syntactic knowledge
 it is system-dependent
 different

 keyboards, commands, function keys, sequences of actions

 some overlap may exist (e.g. with arithmetical operations)

 s for sending a message - s for saving a file - ...
 to overcome these problems, new interfaces show familiar objects and actions
representing the user’s task objects and actions
 standard widgets are easily available
Principle 1: recognize diversity
 human diversity X, different situations X, different tasks X...the set of design
possibilities increases enormously
 a kid playing a videogame is quite different from a librarian managing books, journals
and documents
 a student navigating the net for renting a surf board and locating a windy lake/sea is
a long way from a clerk performing hotel reservations many hours a day
 user differences: background knowledge, training in the use of the system, frequency
of use, goals, impact of a user error: impossible to cope with all these issues in one
single design

Usage profiles
 "Know thy user" Hansen (1971): simple, difficult and undervalued goal
 many designers assume they understand both the user and his tasks
 some users prefer to deal with tables, others with graphs, with words instead of
numbers, etc.
 all design should begin with an understanding of the intended users including
  cultural or ethnic background, education, training, motivation, goals, age, gender,
physical abilities, personality

Different user communities

 the same system may be used by different communities
 high school teachers, nurses, doctors, computer programmers, librarians have
various combinations of knowledge and usage patterns
 other variables: urban-rural, rich-poor, able-disable, attitudes to technology
(explorative-conservative)
 skills at comprehension of Boolean expressions, knowledge of set theory, fluency in
a foreign language, skills in human relationships

Different users
 it is a never-ending process
 users evolve, new users arrive
 novice:know little about the task or interface
 first time users:know little about the interface
 designers must cope with both:clear instructions, help, manuals, tutorials,...
 knowledgeable intermittent users

 they generally use many different systems

 stable task concepts broad knowledge of interface concepts
 difficulty in retaining the menu structure or feature location

 the designer will provide an orderly menu structure, consistent terminology, high
interface apparence (more recognition, less recall), on line help, protection from
dangerous situations,...
 expert frequent users:

 very familiar with the task & interface concepts

 want their work done quickly
 rapid response times, brief and non distracting feedback, complex actions carried
out with few, simple commands (short cuts)

 the designer will provide macro facilities, abbreviations, accelerators

Multi user classes

 these systems must have a level-structured approach to learning
 training-wheels interface for novices

 using examples
  testing their reached ability

 this approach must be also contained in the manuals, help screens, error messages
and tutorials
 interaction pace (slow for novices and fast for frequent users)

Task profiles
 once the user profile is determined, tasks should be focused
 this analysis should precede the design process
 design or implementation convenience should not dictate system functionality or
command features
 high-level tasks first but, until which level should be go to define low-level, atomic,
tasks?

 too small: too many

 too large: require special options

 relative task frequencies are important in shaping a menu tree

Text editor, an example

 frequent actions via special keys (4 cursor arrows, insert, delete)
 intermediately frequent actions (single letter + ctrl, or a selection from a pull-down
menu) like underscore, center, indent, subscript, superscript
 infrequent actions will require menu selections, form fillins like when changing
printing format or revise network-protocol parameters

Medical inf. system: frequency of task

Query by Update Query across Add Evaluate

Job Tilte
patient Date patient Relations System
Nurse 0.14 0.11
Physician 0.06 0.04
Supervisor 0.01 0.01 0.04
Appointment
0.26
personnel
Medical-record
0.7 0.04 0.04 0.01
mainteiner
Clinical research 0.08
Database
0.02 0.02 0.05
programmer
 Interaction styles
 menu selection
 form fillin
 command language
 natural language
 direct manipulation

Direct manipulation

ADVANTAGES DISADVANTAGES
visually presents task may be hard to program
allows easy learning may require graphics
allows easy retention device
allows errors to be avoided
encourages exploration
affords high subjective satisfaction

Menu selection

ADVANTAGES DISADVANTAGES
shortens learning presents danger of many menus
reduces keystrokes may slow frequent users
structures decision making consumes screen space
permits use of dialog-management
tools requires rapid display rate
allows easy suport of error
handling

Form fillin

ADVANTAGES DISADVANTAGES
simplifies data entry consumes screen space
 requires modest training
gives convenient assistance
permits use of form-management tools

Natural language

ADVANTAGES DISADVANTAGES
relieves burden of requires clarification learning syntax
dialog may require
more keystrokes
is unpredictable
may not showcontext

Interface Design
8 golden rules
1 strive for consistency
most violated rule since there are many forms of consistency

 similar situations should require similar action sequences

 identical terminology
 consistent colour, layout, capitalization, fonts
 exceptions: no echo for password or confirmation of delete command should be
comprehensible

2 enable frequent users to use shortcuts

as frequency increases the users desire to reduce the number of interactions to
increase speed

 abbreviations, special keys, hidden commands, macro facilities

 short response times
 fast display rates

3 offer informative feedback

for every user action there should be an appropriate feedback

 for frequent and minor actions, response can be modest

 for infrequent and major actions, response should be substantial
 visual presentation of objects of interest are convenient ways to show changes
  visualize the system state

4 design dialogs to yield closure

sequences of actions should be organized to provide beginning, middle and end

 informative feedback at the completion of a group of actions provides satisfaction

to the user, a sense of accomplishment, relief
 it is the signal to drop contingency plans and options
 the way is clear for the next group of actions

5 offer error prevention and simple error handling

design systems where users cannot make mistakes

 prefer menu selection to form fillin

 do not allow alphabetic characters in numeric entry fields
 if an error is made the system should detect the error
 offer simple, constructive and specific instructions for recovery
 users should only retype the faulty part of a command

6 Permit easy reversal of actions

permit easy reversal of actions

 actions should be reversible as much as possible

 exploration is encouraged, anxiety is diminished
 different units of reversibility: an action, a data entry task, a name, an address
block

7 Support internal locus of control

support internal locus of control

 users want to feel in charge of the system

 no surprising responses, inability to obtain necessary information, inability to
produce the wanted response
 avoid a-causality
 make the user an action initiator instead of a responder to system actions

8 Reduce short term memory load

reduce short term memory load

 users may remember 7+ objects, 2-info chunks

 displays should be kept simple
 window motion frequency reduced
 sufficient training time allotted for codes, mnemonics, action sequences
 online access to command-syntax forms, abbreviations, codes should be
provided
 To sum up
 The above principles should be interpreted, refined, extended and integrated for
each environment
 productivity may be increased by simplifying data entry procedures, making the
displays easy to interpret, providing fast feedback
 all the above will enhance the feeling of control of the user

On errors
 No rule has been found against human error (Sigmund Freud)
 users of computer programs make a lot of mistakes! more than expected
 Card reports that experienced professional users made mistakes or used inefficient
strategies in 31% of the tasks assigned to them
 one way to reduce mistakes is to provide good error messages
 Shneiderman reported that, with improved error messages produced success at
repairing the errors, lower error rates, increased subjective satisfaction

Error messages
 better messages mean:

 more specific: unmatched left parenthesis

 positive in tone: menu choices between 1 and 5
 constructive: try (a sequence of actions)

 instead of

 vague and hostile: 'syntax error'

 intimidating: program aborted, illegal data, system crash
 too technical: postscript error, stack overflow

Error prevention
 this goal is more attainable than what it seems
 understand the nature of errors
 Norman’s view is that people make 'slips' and designers help them to properly
recognize commands, guide them to avoid irreversible actions, make menu choices
distinctive
 Norman’s advice is:

 to avoid modes
  to offer feedback about the system state
 to make commands consistent

Error reduction
3 techniques for reducing errors:
correct matching pairs
complete sequences
correct commands
1
typical example is the parenthesis completion ((())) which can be easily spotted by the
system and reported to the user.
Other markers (<<<>>>) to delimit boldface, italic or underscored text in word
processors or for web programming (html language).
To prevent this, an editor could put both markers and the users fills in avoiding his last
marker slip; for some users this approach is too rigid, a message for 'missing marker'
could be placed in a screen corner
2
if an action requires a sequence of commands, a single command may condense the
full sequence

 a phone dialing up procedure requiring a single button press

 loggin on, style in a word processor
 loading files, etc.

at the same time, users should be allowed to modify the sequence of actions and have
atomic control
both macro action condensation and free single actions should be allowed

Last strategy: correct commands

 annoying typographic errors, pressing more than one key, misspelling a file name,
omitting, inserting or transposing characters

 invoke commands which are not available

 request files that do not exist
 enter data values out of range

 error messages

 ? / SYNTAX ERROR / ILLEGAL COMMAND

 some systems offer:

  automatic command completion
 direct manipulation for choosing the correct action pointing at one of the possible
(correct) choices

based on:

 ample screen space

 rapid display rate
 pointing device is fast and accurate

Guidelines for data display

 importance of guidelines

 i/o formats
 terminology
 command sequences
 terminology
 h/w devices

 system analysts + graphic designers + programmers + HCI experts

Organizing the display

Smith & Mosier (1986)
There are five level objectives for data display
1

 consistency of data display:

terminology, abbreviations, formats, colors, capitalization...all standard.
2

 efficient information assimilation by the user:

familiar format, application domain dependent,rules for neat data columns, left
justification for alphanumeric data, right justification for integers, lining up
decimal points, proper spacing, comprehensible labels,...
3

 minimal memory load on user

not necessary to remember contents of one screen to relate them to another
screen;
completion of tasks should be made by few actions;
labels and common formats should be provided for novices - intermittent users
4
 compatibility of data display with data entry
format of displayed information clearly linked to the format of data entry;
o/p fields should also act as editable i/p fields (debatable: in DBMS users get
confused between query and answer)

 flexibility for user control of data display

the user should be able to chose the best way to view the information on the screen;
typically, the order of columns and rows should be easy to modify.

Extensions
 the above guidelines are useful but need to be extended and adapted to the
application domain and to the computing platform
 from a report on the design of control rooms for electric power utilities - Lockheed
(1981)

 consistent labeling & graphic conventions

 standardize abbreviations
 consistent format in all displays: headers, footers, paging, menus,...
 present a page number on each display page, allow actions to call up a page via
entry of a page number
 present data only if they assist the operator
 present information graphically, where appropriate, by using widths of lines,
positions of markers on scales, and other techniques that relieve the need to
read and interpret alphanumeric data
 present digital values only when knowledge of numerical value is necessary and
useful
 use high-resolution monitors and maintain them to provide maximum display
quality
 design a display in monochromatic form using spacing and arrangement for
organization and then judiciously add color where it will aid the operator
 involve users in the development of new displays and procedures

 many of these issues have already been covered

 the main lesson to be learned is that only relevant information (for the task) should
be displayed
 the visualization should be human-oriented

User’s attention
 exceptional conditions or time-dependent information should become immediately
visible by the user through
Intensity
two levels only
Marking

underline
bold
italic
Size

up to four sizes larger sizes attract attention: A B C D

Choice of fonts

use up to 3 fonts: Algerian, Arial, Times

Inverse video

Inverse coloring
Blinking

(2/4 hertz) with great care and in limited areas

Colour

up to four standard colours, with additional colors for occasional use

Color blinking

changes in colour with great care

Audio

soft tones for regular + feedback and harsh sounds for emergency

Caution
 dange of cluttered displays by overusing the above techniques
 novices need simple, logically organized and well labeled displays
 expert users do not need extensive labels on the fields but subtle highlighting or
positional presentation is sufficient
 display formats should be tested for comprehensibility
 similarly highlighted items will be perceived as being related
 colour coding is powerful in linking related items, difficult to cluster items
 the user should be allowed to highlight as wished

Audio highlighting
 audio tones signal feedback about progress - clicks in keyboards/ringing sounds in
telephones
 alarms for emergency conditions do alert users rapidly but a mechanism to suppress
alarms must also be provided
 different audio alarms must be tested for user recognition
 alternatives are pre-recorded messages (or synthesized) but may interfere among
different operators

Data entry: guidelines

 typical source of errors
Smith & Mosier (1986)
 Five level objectives - Smith & Mosier (1986)
1
 consistency of data-entry transactions

 similar delimiters, abbreviations,...

2
 minimal i/p actions by user

 fewer i/p actions imply greater productivity

 fewer chances of error
 better a single choice (mouse click, menu selection, single key) than a sequence
of characters
 only 1 device better than switching between mouse, keyboard, light pen...where
home row position is lost
 redundant data entry should be avoided

it is annoying to provide the same information in different places twice,

opportunity for error - (the system should copy the information)

3
 minimal memory load on users
users should not need to remember lengthy codes and complex syntactic
command strings
4
 compatibility of data entry with data display
closely link format of displayed information with data-entry format
5
 flexibility for user control of data entry
experienced operators may prefer a given entry sequence, this clashes with the
consistency principle

Automation vs human control

 up to now we have tried to accommodate the user performing his task - (computers:
closed systems)
 users should avoid working through routines, tedious procedures, error-prone tasks
so that:
they concentrate on critical decisions, planning and coping with unexpected
situations

human judgement is necessary for unpredictable events (world: open system) in

which actions must be taken to preserve safety, avoid expensive failures,
increase product quality - Hancock and Salien (1996)

Humans and machines: capabilites

Brown, 1988 - Sanders and MacCormick, 1993

HUMANS GENERALLY BETTER MACHINES GENERALLY BETTER

sense low level stimuli sense stimuli outside human’s range
detect stimuli in noisy background count or measure physical quantities
recognize constant patterns in varying
store quantities of coded
situations
monitor prespecified events especially
sense unusual and unexpected events
infrequent ones
make rapid and consistent responses to i/p
remember principles and strategies
signals
retrieve pertinent details without a priori
recall quantities of detailed info accurately
connection
draw on experience and adapt decisions to process quantitative data in prespecified
situation ways
reason deductively: infer from general
select alternatives if original approach falls
principle info accurately
reason inductively: generalize from
perform repetititve actions reliably
observation preprogrammed
act in unanticipated emergencies and novel
exert great, highly controlled physical force
situations
apply principles to solve varied problems perform several activities simultaneously
make subjective evaluations simultaneously perform several activities
maintain operations under heavy
develop new solutions
information load
concentrate on important tasks
adapt physical response to changes in maintain performance over extended period
situation of time
 Decision making
 with air traffic control, unexpected situations might occur

 a plane may request immediate landing (a sick passenger)

 another plane has a failed engine
 who should land first?

 the human should take responsibility and decide, no automatic procedure can deal
with such situations
 another emergency: fire on a plane, transponder out of service and no
communication between pilot and air traffic control tower...

System design
 goal: to give operators sufficient information about current status and activities
 when intervention is necessary, they hae the knowledge and capacity to perform
correctly even under partial failures
 human role is to respond in unanticipated situations:

 equipment failure
 improper human performance
 incomplete or incorrect data

On automation
 for both home & office automation, some advocate agents (antropomorphic,
adaptive, autonomous) carrying out the user’s intents and even anticipate their needs -
Hayes-Roth (1995)
 Aple’s helpful young man, HPackard’s Bill, Microsoft’s Bob, etc. as butlers, guides,
on-screen partners
 these characters are called knowbots orsoftbots since they are knowing robots that
or programs that know
 human-human interaction may be a good model for HCI, there is a tendency to
design computer partners with antropomorphic (or credible) representations

Helpful avatars
 so far, the animated characters placed at information kiosks have proved unhelpful
 this is true for many other tentative solutions to the information-provider agent (bank
assistants, car driver informer, automatic postal clerk, etc.)
 a variant to the anthropomorphic agent is a program, based on a user model, to
guide an adaptive system which modifies the interface making life easier for the user.
Adaptations based on:
response time - length of messages
density of feedback- order of menu items
i.e. for a frequently used item in a menu, such item should come first

Dark side
 if the adaptive system makes a change and it is unexpected, the user must pause,
understand and resume his work - time lost...
 the system might ask the user if he accepts the change but then, again, we have job
disruption
 designers believe that users are attracted by "autonomous, adaptive, intelligent"
systems
yet
 users may feel anxious and out of control
 only advantageous systems have been for training & help in well studied situations

Another path
 the negative issue has encouraged another approach: www searching and
collaborative filtering
 no visible agent or adaptation in the interface: the applications aggregate information
from multiple sources in some (often proprietary) way
 black box approach which are entertaining like when selecting movies, books,
music...
 sometimes searches are to broad and too many items are captured

Alternative to agents
 user-control, responsability and accomplishment
 users desire to be in full control and gain system mastery
 comprehensible and predictable user interfaces should mask the underlying
computational complexity
 critical issue: who is responsable for failures?
 an alternative may be to expand the control panel metaphor
now: set up for speed of cursor blinking, rate of mouse tracking, speaker
loudness, personal preferences,...
next: set up for different user levels, complex processes are specified by direct
manipulation programming, graphical specifications,...

Control Panels for Mac

 Control panels for Windows

Summing up
 designing user interfaces is complex & creative
 task analysis should come first and be detailed and accurate
 recording of task objects and actions leads to useful metaphors for interface objects
and actions
 principles and guidelines are emerging from practical experience
 guidelines support consistency, records organizational policies, facilitates training of
new designers, stimulates discussion
http://www.aw.com/DTUI
 The design process
 old users where available to learn difficult programs that would fulfill their needs
 present users are experts in their work domain but not technicians
 design should address their issues
 1991 - the Usability Professionals Association was born
 LUCID - (Logical User-Centered Interaction Design) includes the observation of
users, the development of real scenarios and of social impacts coming from using new
systems

Usability support
 CUO - Chief Usability Officer, Usability Day, Usability Managers...
 to stimulate change

 note the difficulties of users, their mistakes...

 evaluate peformance data on interfaces developed with LUCID

 each project has its architect who develops the user interface and coordinates

 human factors specialists

 reads the technical literature,prepares budgets and deadllines
 designs usability tests

Specialization
 today’s jobs narrow - we have specialists in:

 interface development tools

 data visualization strategies
 use of sounds (voice synthesis, speech recognition)
 technical writing, handbooks...

Development tools
 choosing the right tool is one basic step
 new instruments speed up the implementation and favour the desig-test-refinement
cycle
 the writing of the guideline document is only the first step towards the implementation
 management strategies, also called 3E (enforcement, exemption, enhancement) are
spreading around
 Still on usability
 it is always difficult to grasp the concept of usability (every change is costly)
 it is also difficult to include the user’s requests in the project
 Karat (1990, 1994) noted that for every single dollar spent on usability,100 US$
payoffs were obtained spread out along the following items:
 reduced program development cost

 reduced program maintenance cost

 earning increse due to greater user satisfaction
 efficiency and produttivity increase

Other data
 Landauer (1995) showed productivity increases up to 720% in projects that
considered usability from the start
 even minimal modifications of a project, after a usability test which pinpointed 20 of
the most common errors, increased user efficiency from 19% up to a maximum of 80%
 usability engineers become the real project managers, they also take care of
resource allocation, costs and deadlines

Design
 design is typically creative and unpredictable
 Carroll & Rosson (1985) describe design as follows:

 design is a process, not being a state it cannot be described statically

 design is non hierarchical,neither bottom-up nor top-down
 deisgn is radically transformational; it requires the development of partial and
temporal soutions which may also be eliminated during the overall design
process
 design also includes the discovery of new goals

The three Pillars of Design

 3 pillars of design, each one produces a time gain of one order of magnitude
 moreover, the system quality is improved
 First pillar
 a set of working guidelines should be available from the start
 10 page document obtained from one week work from 2 persons
 Macintosh was a good example of a generated guideline for programmers that easily
read and understood what they had to do to maintain compatibility
 Microsoft also has a document of this sort for implementors

Guidelines
 Words and icons

 terminology (objects and actions), abbreviations, and capitalization

 character set, fonts, font sizes, styles (bold, italic, underline)
 icons, graphics, and line thickness
 use of colour, backgrounds, highlighting, and blinking

 Screen-layout issues

 menu selection, form fillin, and dialog-box formats

 wording of prompts, feedback, and error messages, justification, whitespace, and
margins
 data entry and display formats for items and lists
 use and contents of headers and footers

 Input and output devices

  keyboard, display, cursor control and pointing devices
 audible sounds, voice feedback, touch input, and other special input modes or
devices
 response times for a variety of tasks

 Action sequences

 direct manipulation clicking, dragging, dropping, and gestures

 command syntax, semantics, and sequences
 programmed function keys
 error handling and recovery procedures

 Training

 online help and tutorials

 training and reference materials

Writing the guidelines

 it is a social process to gain visibility and build support within an organization
 controversial guidelines should be reviezed by colleagues or tesed
 procedures to distribute the guidelines
 ensure enforcement
 allow exemptionspermit enhancements (the 3Es)
 guidelines are a living text adapted to changing needs and refined through
experience
 acceptance may be increased by rigid standards, accepted practices and flexible
guidelines

How to use the guidelines

 if properly written and provided at the beginning of the project, they may become an
opportunity for discussion of controversial issues
 implementation proceeds quickly and with few design changes
 they announce policies for

 enforcement - who reviews?

 exemption - who decides?
 enhancement - how often?

User-interface s/w tools

 both users and even designers have a vague idea of what the system will look like
when built
 once the system has been developed, major changes involve high cost and time
 a realistic impression of the future system must be provided to the users

 printed version of the proposed displays

 on-screen display with an active keyboard & mouse
 prototype of a menu may have only a few active paths...

Prototypes
 a form-fillin system may show the fields without processing them
 paper models are also helpful
 Hypercard and Macromind Director are good building tools for prototypes
 Visual Basic and Borland’s Delphi are alos used
 Visix’s Galaxy and Sun’s Java are more sophisticated

Expert reviews and usability testing

 theatrical producers know that previews to critics and extensive rehearsals are
necessary for a good opening night
 dress rehearsals are necessary with full cast, props and lighting
 interactive system designers must carry out many small an some large pilot tests
 expert review methods, tests with intended users, surveys, automated analysis tools,
are all required

Development methodologies
 typical failure of s/w projects is of 60% !!

 25% never completed

 35% achieving partial success

 due to lack of attention to design issues at the start

 well designed systems are less expensive to develop and maintain
 easier to learn and use and provide the users with a sense of mastery and
confidence encouraging exploration

Design in practice
 formal methodologies have relieved the pain of deadlines and budgets for
developers
 usable interfaces are still something different...advocated by academics
 improved methodologies for new hi-tech companies which have built from the original
user-centered approach
 detailed deliverables are specified incorporating cost/benefit and return on
investment (ROI) analysis
 moreover, management strategies - for schedule and budget - are also considered

Logical user-centered design methodology: LUCID

 formerly called QUE for quality Usability Engineering (Kreitzberg, 1996)
 it identifies 6 differents stages

1. develop product concept

2. perform research and needs analysis
3. design concepts and key-screen prototype
4. do iterative design and refinement
5. implement software
6. provide rollout support

Stage
1
 create a high concept
 establish business objectives
 set up the usability design team
 identify the user population
 identify technical and environmental issues
 produce a staffing plan, schedule and budget

Example of concept: The new home banking system will provide customers with
unified access to their accounts. It will support balance inquiry, management of
credit acounts and loans, transfer of funds, among acounts, electronic bill
payment and investment in the bank’s family of mutual funds. The system will
provide the custormer with year-end accounting for tax purposes.
2
 partition the user population into homogeneous segments
 break job activities into task units
 conduct needs analysis through construction of scenarios and participatory design
 sketch the process flow for sequences of tasks
 identify major objects and structures which will be used in the software interface
 research and resolve technical issues and other constraints
3
 create specific usability objectives based on user needs
 initiate the guidelines and style guide
 select a navigational model and a design metaphor
 identify the set of key screens: login, home, major processes
 develop a prototype of the key screens using a rapid prototyping tool
 conduct initial reviews and usability tests
4
 expand key-screen prototype into full system
 conduct heuristic and expert reviews
 conduct full-scale usability tests
 deliver prototype and specification
5
 develop standard practices
 manage late stage change
 develop online help, documentation and tutorials
6
 provide training and assistance
 perform logging, evaluation and maintenance

Key screen prototype

 one important feature of LUCID is the key screen prototype which - in participatory
design sesions - is looked at with critical views conveying the system concept to the
non-technical users
 LUCID solicits user i/p, construct workflow scenarios and define the objects central to
the design
 the key screen prototype incorporates the major naavigational paths onf the system
allowing the users to evaluate and refine it
 it is also used for usability testing and heuristic review
 key screens evoke strong reactions, generate early participation, create momentum
for the project

On LUCID
 based on rapid prototyping, iterative usability testing
 spots interface issues with have implications for the technical architecture of the
product
 LUCID describes a phased rollout approach built on theories of organizational
change
 LUCID makes committments to user-centered design and highlights the role of
usability engineering in s/w development focusing on:
 activities

 deliverables
 reviews

Areas of activity
 For each stage, 9 areas of activity are evaluated in connection with specified
deliverables and timely feedback through reviews

1. product definition
2. business case
3. resources
4. phys. environment
5. technical environment
6. users
7. functionality
8. prototype
9. usability
10. design guidelines
11. content materials
12. documentation training and help

LUCID conclusions
 each project is different yet a design methodology is mandatory
 validation & refinement in multiple projects has proven useful & constructive
 LUCID promotes an orderly process and predictable progress

Ethnographic observation
 observation of users
 an ethnographer participate - overtly or covertly - in peole’s daily lives for an
extended period of time watching what happens, listening to what is said, asking
questions
 knowledge on the organizational context and individual behaviour is gained
 user-interface designers observe the users AND the interfaces they use at work
 etnographic methods have been applied to office work, air-traffic control, other
domains

Etnographic guidelines
 preparation

 understand organization policies and work culture

 familiarize yourself with the system and its history
 gain access and permission to observe or interview

 field study

 establish rapport with managers and users

  observe or interview users in their workplace, and collect subjective and objective
quantitative and qualitative data
 follow any leads that emerge from the visits

 analysis

 compile the collected data in numerical, textual, and multimedia databases

 quantify data and compile statistics
 reduce and interpret the data
 refine the goals and the process used

 reporting

 consider multiple audiences and goals

 prepare a report and present the findings

Comments
 each collected data requires interpretation and attention for each situation
 learning the technical language of the users is vital for obtaining rapport
 rating scales or rankings should be used
 anecdotes or critical incidents capture user experiences
 written report summaries are also useful
 working relationships may develop
 users become active participants in the design of their new interface

Participatory design
 Positive features

 more accurate information about taks

 an opportunity for users to influence design decisions
 users’ ego investment makes for successful implementation
 increased user acceptance

 Negative features

 costly and lengthy

 builds antagonism with people not involved
 force designers to compromise
 builds opposition to implementation
 PICTIVE
 a Plastic Interface for Collaborative Technology Initiatives through Video Exploration
(Muller, 1992)
 users sketch interfaces, use slips of paper, pieces of plastic and tape to create early
prototypes
 a scenario is video recorded for presentation to managers, users, other designers
 PICTIVE elicits new ideas and be fun

Project leader
 considers social and political environments
 the right level of user involvement
 personalities of the users and design team members
 individual preferences and organizational politics are sometimes stronger than
technical issues
 novelty is threatening to many people...
 clear statements about what to expect when should be produced

Scenario development
 for current interfaces data on the range and distribution of task frequencies and
sequences are available
 if not available they should be logged from day to day operation
 these data are crucial, particularly if a process known as business-process-re-
engineering is taking place

 where the ts are single tasks and the fs are single frequencies for every user class

Preparing the field

 data about current performance should be collected
 information about similar systems can be gathered
 interviews can be conducted with interested parties (Carroll, 1995)
 an early and easy way to describe a novel system is to write scenarios of usage and
then act them out
 when multiple users must cooperate as in
  cockpits, control rooms, financial trading rooms this technique is particularly
useful

 or using multiple devices as in

 customer-servie desks, medical laboratories,...

National Digital Library Project

 an example of scenario portraying the NDL Project
K-16 Users: a seventh-grade social-studies teacher is teaching a unit on the
Industrial Revolution. He wants to make use of primary source material that
would ilustrate the factors that facilitated industriallization, the manner in which it
occurred, and the impact that it had on society and on the built environment.
Given his teaching load, he only has about four hours total to locate and package
the supplementary material for classroom use.

Another scenario
 US Holocaust Museum and Education Center
A grandmother and her 10-and 12- year old grandsons have visited the museum
before. They have returned this time to the Learning Center to explore what life
was like inn her shetl in Poland in the 1930s. One grandson eagerly touches the
buttons on the welcome screen, and they watch the 45-second video introduction
by the museum director. They then select the button on “History before the
Holocaust” and choose to view a list of towns. Her small town is not on the list,
but she identifies the larger nearby city, and they get a brief textual description, a
map of the region, and a photograph of the marketplace.

Holocaust Museum and Education Center

They read about the history of the town and view 15-second videos of the market-place
activity and a Yiddish theater production. They bypass descriptions of key buildings and
institutions, choosing instead to read biographies of a famous community leader and a
poet. Finally, they select “GuestBook” and add theirnames to the list of people who have
indicated an affiliation with this town. Further up on the list, the grandmother notices the
name of a childhood friend from whom she has not heard in 60 years-fortunately, the
earlier visitor has left an address.

Analysis of scenarios
 written to give nontechnical museum planners and Board of Directors the idea of
what could be built if funding were provided
 easy for people to grasp
 convey design issues such as physical installation and development requirements
 some scenario writers also produce video tapes, e.g.

 Apple’s Knowledge Navigator (1988)

 Sun’s Starfire (large screen work supporting CSCW)
 Bill Gates police drama 2005 (digital wallets, interactive home TV, educational
dbs, medical communications

Social impact
 issues of technology - web, privacy, copyright, etc.
 impact of new systems requires a statement (as in ecological policies) to promote
high quality systems in government related applications - Battle et al., (1994)
 in the private sector reviews would be optional and self administered
 an outline was suggested by Shneiderman and Rose (1996)

Social impact statement

 describe the new system and its benefits

 convey the high-level goals of the new system

 identify the stakeholders
 identify specific benefits

 address concerns and potential barriers

 anticipate changes in job functions and potential layoffs

 address security and privacy issues
 discuss accountability and responsibility for system misuse and failure

Missing part of statement

 avoid potential biases
 weigh individual rights versus societal benefits
 assess tradeoffs between centralization and decentralization
 preerve democratic principles
 ensure diverse access
 promote simplicity and preserve what works

Final part of statement

 outline the development process
  present an estimated project schedule
 propose process for making decisions
 discuss expectations of how stekeholders will be involved
 recognize needs for more staff, training, and hardware
 propose plan for backups of data and equipment
 outline plan for migrating to the new system
 describe plan for measuring the success of the new system

Social impact
 written by the design team
 early along the process design
 accessible to users, managers, public
 the statement is next evaluated by managers, other designers, end users, federal
government units, stte legislatures, regulatory agencies, professional societies, and
labor unions
 once the social impact statement is adopted it must be enforced
 effort, cost and time should be appropriate to the project

Statement consequences
 it may offer large improvements by preventing problems that could be expensive to
repair
 improves privacy protection
 minimizes legal chalenges
 creates more satisfying work environments
 the common goal is excellence in design

Legal Issues
 every development process should include a review of legal issues that may affect
design, implementation or marketing
 privacy: medicl, legal, financial, military...
 illegal tampering, inadvertent loss, malicious mischief
 privacy protection may involve user-interface mechanisms for controlling passwords,
file access, identity checking, data verification
 encryption and decryption processes may involve complex dialog boxes to specific
keys
 safety and reliability:

 aircraft, automobiles, medical equipment, military systems, nuclear reactor

control rooms...may require life-or-death decisions like in air-traffic control
 if an interface is difficult to understand it could lead to
 a law suit against the designer, developer, implementor alleging improper design
 high-quality interfaces, well-tested, adhering to state-of-the-art design guidelines
 copyright protection for software and information
 protection against piracy
 clever hackers usually circumvent all barriers
 sues against corporations and universities are not uncommon
 site-license agreements allow copying within the site
 new problems arise with on-line information (such as the one from a database or
from the web)

 who owns it? can one sell it?

 what about publishers and authors?

Another issue
 freedom of speech in electronic environments
 are users allowed to make controversial/potentially offensive statements via e-mail?
 must community standards hold?
 are ntework operators responsible for or prohibited from eleiminating offensive or
obscene jokes, stories or images?
 equal access for impaired persons
 the most controversial issue for user-interface designers is: copyright and patent
protection for interfaces

What should be protected?

 what material should have copyright?

 fonts, lines, boxes, shading and colors cannot be accorded copyrights i.e. GUIs
cannot be protected

 music is protected, notes cannot be (collections of words may be copyrightable)

 yet...Apple’s desktop was copied by Microsoft (the judge decomposed the interface
into elements)
 the confusion lies on the difficult separation between ideas (not protectable) and
espressions (protectable)

An example
 the idea of working with many documents at once is not protectable (by multiple
windows) but that specific expressions of windows (colored frames, animation, etc.) is
protectable
 are copyrights or patents more appropriate for user interfaces?
  copyright is easy to obtain - a notice on the user interface and file an application -
last 75 years for a company and life +50 years for persons
 patent is complex, slow and costly - last 17 years but are more enforceable

Copyright infringement
 it is easy to detect if the copy is complete
 what if there are partial copy, some elements are the same but not all of them
 ordinary observers must be convinced that the copy is substantially similar to the
original

User interfaces copyrighted?

 some believe that user interfaces should not be copyrighted, instead they should be
shared
 copyrights interfere with standardization
 others believe that user interfaces should be copyritten for recognizing artistic work,
encourage innovation and reward the designers
 complexity and cost of protection, designers do not want to share their creatures
 interface designers must respect existing expressions, they should seek for licenses
or cooperative agreements to share user interfaces

Evaluating the designed system

The elements that concur to the system evaluation are:

 stage of design early, middle, late

 novelty of project well defined vs exploratory
 number of expected users
 criticality of the interface:life-critical medical system vs museum-exhibit support
 costs of product and finances allocated for testing
 time available
 experience of the design and evaluation team

Ranges
 the evaluation plans might be a 2 years test with mutiple phases for an air-traffic-
control system or
 a 3 day test with 6 users for a small internal accounting system
 costs may vary from 10% of the project to 1%
 not possible now to bypass usability testing
 customers might file lawsuits to s/w vendors for errors
 Limitations
 impossible to test the system in every possible situation
 testing must include continuing methods to assess and repair problems during the
lifecycle of the interface
 a decision must be made as to delivery after testing
 most testing methods account for normal usage but stressful situations and partial
equipment failures should also be considered
 more than 4000 members of the Usab.Prof. Ass. exchange ideas about these
problems

Expert reviews
 experts may belong to staff or be external consultants
 reviews may be conducted rapidly
 reviews may be performed early or late in the design phase: they provide
recommendations, a formal report or both
 suggestions should be made cautiously (take care of the designer’s ego)
 it is better to pinpoint problems than to provide solutions
 solutions should be left to the designers

Expert review: methods

 heuristic evaluation
i.e. evaluate using the 8 golden rules; expertise on the rules is very important
 guidelines review
the interface is checked with respect to organizational & guidelines documents
 consistency inspection
across a family of interfaces: color, layout, terminology, i/p & o/p formats, training
materials, help
 cognitive walkthrough
users are simulated in carrying out their tasks; (Wharton, 1944) frequent tasks
are the starting point, critical tasks, error recovery, also public walkthroughs are
performed (Yourdon, 1989)
formal usability inspection
courtroom-style meetings with a moderator to present the interface and discuss
merits and waknesses, design team member may rebut; these meetings may be
good experiences for managers yet they are time consuming.

Expert review vs usability studies

 difficult to compare, different contributions to improve the interface
 some studies prove the benefits of expert reviews (Jeffries et al 1991, Karat et al
1992)
 different experts find different problems, it may be a good thing to use 3-5 experts
and collect all evidence
 expert reviewers should act in the same conditions as the potential users (work
place, noise, stress)
 bird’s eye view of an interface via printed screens pinned to a board may be very
useful to detect inconsistencies
 some experts may lack knowledge on the task domain, conflicting advice may be
negative
 experts should have a long term relationship with the organization - they may be
accountable
 difficult to predict how first time users will behave

Usability testing and laboratories

 started in 1980
 traditional managers resisted (nice idea but time and money pressures prevented
them from adopting usability evaluation)
 competition started the need for such evaluation, moreover deadlines could be met if
an usability test was scheduled
 the results of the test provided:

 supportive confirmation of progress

 specific recommendations for changes

 designers had evaluative feedback to guide their work

 managers saw fewer disasters as delivery dates approached
 usability testing speeded up many projects
 it also produced dramatic cost savings
 usability laboratory tests were influenced by marketing and advertising, few users,
quick & dirty
 controlled experiments tested hypothesis, support theories and methods and
produce statistically significant results

Usability labs
 have emerged in different companies
 they provide a positive image of the company
 some are very large (16 labs at Boca Raton, IBM)
 usability consultancy firms have started and may be hired
 each lab may serve 10 to 15 projects a year; lab staff meets with the user interface
architect or manager at the kick off to make a test plan with scheduled dates and budget
allocations

Pre-test
 usability staff participate in early task analysis, provide info on s/w tools, references
and help develop set of taks for the usability test
 2 to 6 weeks before the test a detailed test plan is defined, inlcuding a list of tasksm
subjective satisfaction and debriefing questions
 number of participants, source: customer site, personnel agency
 a pilot test of procedures, tasks, questionnaires made 1 week ahead of time

Test
 final procedures are now defined

 participants are chosen to represent the user communities

 attention to background in computing
 experience with the task, motivation, education, ability with the interface
language
 control of eyesight, left versus right handedness, age, gender
 other experimental conditions: time of day, day of week
 physical surroundings, noise, room temperature

Etiquette
 participants should always be treated with respect
 informed that THEY are not tested, the system is
 they will be told what they will be doing & for how long
 participation should always be voluntary (informed consent)
 a typical statement could be the following:

Statement of consent
I have freely volunteered to participate in this experiment
I have been informed in advance what my task(s) will be and what procedures will be
followed
I have been given the opportunity to ask questions and have had my questions
answered to my satisfaction
I am aware that I have the right to withdraw consent and to discontinue participation at
any time, without prejudice to my future treatment
My signature below may be taken as affirmation of all the above statements; it was
given prior to my participation in ths study

Cues for testing

 participants may be encouraged to think aloud
 the tester should be supportive of the participants taking notes and not interfering
 typically tasks will be achieved after 2-3 hours
 participants are invited to make general comments/suggestions
 sometimes 2 users cooperate in the task and exchange ideas
 videotaping is often performed for later review (tedious job)
 logging all user actions is generally performed, perhaps with special programs (The
Observer, Nolde, The Netherlands)
 logging means tracing mousing, typing, reading manuals, screens, etc.
 designers are impressed when they see (on the tape) the users failing or not
achieving what they want
 sometimes users consistently pick up the wrong menu: the position of that menu was
ackward

Testing & correcting

 at each design stage the interface can be refined iteratively
 the improved version can be tested again
 it is important to fix quickly even small flaws (spelling errors, inconsistent layout,...)
 forms of usability testing have been suggested

 discount usability engineering (Nielsen, 1992) which is a quick & dirty approach
(task analysis, prototype development, testing)
 field test user realistic environments - portable usability labs with videotaping &
logging - a variant is to provide users with test versions of new s/w (Microsoft’s
Windows 95 was screened by 400.000 users!)

Other testing strategies

 early usability testing may be performed with mockups of screen displays to assess
user reactions to wording, layout and sequencing
 a test administrator plays the role of the computer by flipping the pages while asking
a participant user to carry out typical tasks
 game designers pioneered the can-you-break-this approach providing teenagers with
the challenge to beat new games
 this last approach is a destructive one which tries to detect fatal flaws and appears
very productive for critical systems



Testing conclusions
 last approach compares different versions of the same interface or with other similar
interfaces intended for the same job
 its name: competitive usability testing
 it is important to construct parallel sets of tasks and counterbalance the order of
presentation of interfaces
 usability has at least 2 serious limitations:

 emphasis on first usage (2 to 4 hours, remaining period has unknown problems)

 limited coverage of interface features (few aspects may be touched on during a
test)

Surveys
 due to the above conclusions, usability tests must be integrated with other
measurements, i.e. with surveys
 surveys are a familiar, inexpensive and generally acceptable companion for usability
tests and expert reviews
 clear goals in advance + focused items helping to attain those goals
 care in administration and data analysis
 it should be prepared, reviewed among colleagues and tested with a small sample of
users
 statistical analyses and presentations should be developed before the final survey is
distributed
 survey goals may be tied to the components of the OAI model of interface design;
subjective impressions about the representation of:
 task domain objects and actions

 interface domain metaphors and action handles

 syntax of inputs and design of displays

 ascertain the user’s

 background - age, gender,origins, education, income

 experience with computers - specific applications, length of time, depth of
knowledge
 job responsibilities - decision making influence, managerial roles, motivation
 personality style - introvert vs extravert, risk taking vs risk averse, early vs late
adopter, systematic vs opportunistic

 ascertain the user’s

 reasons for not using an interface - inadequate services, too complex, too slow
 familiarity with features - printing, macros, shortcuts, tutorials
feelings after using an interface:
confused vs clear
frustrated vs in control
bored vs excited
 Online surveys
 avoid the cost and effort of printing, distributing and collecting paper forms
 many people prefer to answer a short survey displayed on a screen (rather than
filling in and returning a printed form)
 in a survey, a short scale with 5 values was provided

 strongly agree
 agree
 neutral
 disagree
 strongly disagree

Survey with the 5-value scale

 I find the system commands easy to use
 I feel competent with and knowledgeable about the system commands
 When writing a set of system commands for a new application, I am confident that
they will be correct on the first run
 When I get an error message, I find that it is helpful in identifying the problem
 I think that there are too many options and special cases
 I believe that the commands could be substantially simplified
 I have trouble remembering the commands and options and must consult the manual
frequently
 When a problem arises, I ask for assistance from someone who really knows the
system

Results from this survey

 it helps designers to identify problems users are having
 it demonstrates improvement to the interface as changes are made in

 training
 online assistance
 command structures

 progress is demonstrated as subsequent surveys show higher scores

On a text editor usage

 users had to rate the messages from a text editor on a 7 value scale
 Hostile 1234567 Friendly
 Vague 1234567 Specific
 Misleading 1234567 Beneficial
 Discouraging 1234567 Encouraging
 when precise questions are asked, precise answers will be given

Other questionnaires
 Coleman and Williges, 1985 developed a set of opposing features as reactions users
could have with an interface

 pleasing vs irritating
 simple vs complicated
 concise vs redundant

 and then asked to evaluate, on these grounds, a text editor

 another approach is to ask specific questions like:

 readability of characters
 meaningfulness of command names
 helpfulness of error messages

User satisfaction
 the questionnaire for user interaction satisfaction (QUIS) was developed by
Shneiderman and refined by Norman and Chin
 It was based on early versions of the OAI model and covered interface details

 readability of characers
 layout of displays
 interface objects
 meaningfulness of icons
 interface actions
 task issues

Different questionnaires
 the QUIS was developed and is available at Maryland Univ.
 it includes website design and videoconferencing, based on over 50 questions
 it is layered on two levels (general and detailed) and has been undertaken by over a
thousand users
 other scales include the Post-Study System Usability Questionnaire developed by
IBM and the SUS developed by Digital Corp.

Acceptance tests
 explicit acceptance criteria should be established when the requirements document
is written
 user friendly is vague, we rather use:

 time for users to learn specific functions

 speed of tasks performance
 rate of errors by users
 user retention of commands over time
 subjective user satisfaction

An acceptance specification
The subjects will be 35 secretaries hired from an employment agency. They have no
word-processing experience, but have typing skills in the range of 35 to 50 words per
minute. They will be given 45 minutes of training on the basic features. At least 30 of the
35 secretaries should be able to complete, within 30 minutes, 80 percent of the typing
and editing tasks in the enclosed benchmark test correctly.

another specification
After 4 days of regular use of the system, 25 of these 35 secretaries should be able to
carry out, within 20 minutes, the advanced editing tasks in the second benchmark test,
and should make fewer than six errors.
it focuses on performance after regular use
After two weeks, at least 15 of the test subjects should be recalled and should perform
the third benchmark test. In 40 minutes, at least 10 of the subjects should be able to
complete 75% of the tasks correctly.
Dopo due mesi, vengono richiamate 15 delle segretarie prima utilizzate. Gli viene
assegnato il compito di scrivere un brano per completare la fase di sperimentazione. In
40 minuti, almeno 10 segretarie devono essere in grado di completare almeno il 75%
del compito in modo corretto.
iit focuses on retention
still on acceptance testing
 in a large system 8 to 10 tests will be carried out on ° components of the interface,
with different user communities
 other criteria may be also considered:

 subjective satisfaction
 o/p comprensibility
 system response time
 installation procedures
 printed documentation
 graphics appeal
 acceptance context
 if precise acceptance criteria are established, both customer and interface developer
benefit
 user friendliness is avoided (as a standard to comply with)
 contractual fulfillment can be demonstrated
 outside organizations may be more neutral
 central goal of acceptance testing is not to detect flaws but to verify adherence to
requirements

Post acceptance
 after acceptance, field testing is performed before international distribution
 training methods, tutorial material, telephone help procedures, marketing and
publicity can all be improved
 the evolutionary development is forced, in the pre-release phase, through

 early expert reviews

 usability testing
 surveys
 acceptance testing

evaluation during use

 system refinements may take place during use by many different and disseminated
users
 gradual sysem diffusion is useful so that problems may be solved with minimal
disruption
 major changes should be limited (and anticipated) to an annual revision
 stable access to key resources (on the web)
 gradual addition of improvements (patches)

Interviews
 interviews with individual users can be productive since the interviewer may focus on
specific issues of concern
 focus-group discussions are also valuable to generalize comments
 a small fraction of the user community is involved since all contacts are expensive
and time consuming
 yet contact with users leads to:specific & constructive suggestions

example
 a large corporation conducted 45’ interviews with 66 of the 4300 users of an internal
message system
 these interviews revealed that the users were happy with some aspects of the
functionality (the capacity to pick up messages at any site, the legibility of printed
messages and the convenience of after-hours access) but- out of all users -
23.6% had concerns about reliability
20.2% thought that using the system was confusing
18.2% said convenience and accessibility could be improved
16.0% expressed no concerns

 later questions explored specific features

 as a result of this interview, 42 enhancements to the system were proposed and
implemented
 there were also earlier proposals but the interviews changed the set of priorities
so as to better reflect the user’s needs

Continuous data logging

 managers should be able to collect data about the patterns of

 system usage
 speed of user performance
 rate of errors
 frequency of requests for online assistance

 guidance is provided for

 the acquisition of new h/w

 changes in operating procedures,
 improvements to training
 plans for system expansion

Consequence
 the highest fequency error is a candidate for attention

 the message could be rewritten

 training materials should be revised
 s/w could be changed to provide more specific information
 command syntax could be simplified
 staff should examine messages that never appear
 if logging provides data for each command, each help screen, each database record,
then changes to the human-computer interface can be made to simplify access to
frequently used features
 rarely used features should also be analyzed
 at the US Congress legislation, some high-frequency terms like abortion, gun control
and balanced budget could be entered into a list of hot topics
 another advantage of frequently used features detection is that they help in
optimizing performance and reducing costs

Privacy
 logging is positive but may break into personal grounds
 links to specific user names should not be collected
 monitoring individual activity should be known to users (when and what will be done
with the results)
 manager and worker cooperation improves productivity

Online consultant
 provides a very personal assistance
 users feel reassured if they know there is a human available for help
 such consultants are a good source for listing the problems met by users
 some organizations offer a toll-free number for consultancy
 AOL provides live chat rooms for discussion of user problems
users type their questions and get their answers

Getting help
 many groups maintain a standard electronic-mail address of staff@organization that
allows users to get help
 quick help at night from system analysts
 e-mail can also be used to allow users to send messages to the maintainers or
designers
 an online suggestion box encourages to make productive comments

Suggestions
 an online suggestion box invites comments, gets 10 to 20 a day including thoughtful
ones as this:
I find as I get searching through the various Web pages...that I am left with an
unsatisfied feeling. I have been sitting in front of the PC for close to an hour...and have
been stopped and/or slowed due to items that can be directly related to web server
design.
First off, the entry pages are too big and disorganized. Those links that do exist do not
have adequate enough descriptions to direct a user to the information they desire. In
addition, the use of a search engine would greatly facilitate sifting through the
abundance of information that is thrown at the user with any one of these links.

Riguardo i riferimenti
Links should be short, sweet, and specific. Large amounts of material should not be
included in one document on a busy server...
Breaking up these larger documents into smaller, well organized documents may seem
to create an additional burden on programming. However, if intelligence is used in the
creation of such systems, it would not take much.
An Internet directory service for personal names can be found at Knowbot Information
Service with an invitation:
Place a compliment or a complaint in the KIS log file

Electronic bulletin board

 or newsgroups are used by interface designers
 open messages and questions are posted on both and are read/answered by
interested people
 newsgroups cover Human Factors and Ergonomics Society’s Computer Systems
Technical Group
 user newsletters and conferences generally contain:

 information about novel interface facilities

 suggestions for improved productivity
 requests for assistance
 case studies of successful applications
 stories about individual users

controlled psychologically oriented experiments

 techniques for precise measurement
 outline of the scientific method applied to CHI:

 deal with a practical problem and consider the theoretical framework

 state a lucid and testable hypothesis
 identify a small number of independen variables that are to be manipulated
 carefully choose the dependent variables that will be measured

 Judiciously select subjects, and carefully assign subjects to groups

 Control for biasing factores (nonrepresentative sample of subjects or selection of
tasks, inconistent testing procedures
 Apply statistical methods to data analysis
 Resolve the practical problem, refine the theory, and give advice to future
researchers
 the reductionist approach yields narrow but reliable results

Advantages
 managers recognize the power of controlled experiments
 proposals are made for

 new menu structures

 novel cursor control devices
 reorganized display formats

 the experiment may support a decision made providing change in the interface
 fractions of users could be given the "new" version of the interface, performance is
next compared with the "control group"

Measures
 dependent measures can be performance times, user-subjective satisfaction (on a
short scale), error ratesm user retention over time
 novice experimenters are advised to collaborate with experienced social
experimenters
 open questions:

 what techniques can reduce novice user anxiety?

 how can life-critical appplications be tested reliably?
 which is the cheapest and effective way to to evaluate a system from the user’s
point of view?