© firstperson, inc.

floating point, ordering of values, 24 S

for, 25, 26 short, 9
static, 13
G String, 8, 24
garbage collection, 28 strings, 8, 10, 24
goto, 25 super, 17
switch, 25
I synchronize, 16
synchronized, 15
identifiers, 6
if, 25
implements, 11 T
import, 20 this, 11
instanceof, 24, 28 throw, 26
int, 9 transient, 15
integers, 7, 8, 22 try, 26
interface, 18
interfaces, 11, 18 U
Unicode, 6
L characters, 10
length (array length), 10 unprotect, 28
literals, 7
long, 9 V
long long, 9 void, 12
volatile, 14
M
methods, 12 W
while, 26
N
new, 24 X
XOR, logical, 23
O
OR, logical, 23

P
package, 20
packages, 19
postconditions, 21
preconditions, 21
print(), 24
println(), 24
private, 15
protect, 28
protected, 15
public, 15

R
return, 26

38 Oak Language Specification Confidential

Index

Symbols B
!, 23 boolean, 8
–, 23 boolean expressions, 25
!=, 23, 24 break, 25
%, 23 byte, 9
&, 23
&&, 23 C
&=, 23 case, 25
*, 23 casting, 11, 25
*=, 23 catch, 26
+, 23, 24 char, 9, 10
+=, 23 classes, 8, 10, 19, 24
-, unary, 22 clone, 24
–, unary, 23 comments, 6
/, 23 const, 14
/**...*/, 6 constants. See literals
/*...*/, 6 constraints, 21
//, 6 constructors, 16
/=, 23 continue, 26
<, 23, 24
<<, 23
<=, 23, 24
D
–=, 23 declaration order, 17
==, 23, 24 default, 25
>, 23, 24 do, 26
>=, 23, 24 double, 9
>>, 23 double precision, 7, 9, 23
>>>, 23
^, 23 E
^=, 23 else, 25
|, 23 exceptions, 26
|=, 23 cast, 25
||, 23 floating point, 24
~, 22 extends, 11

A F
assertions, 21 final, 15
finally, 27
float, 9
floating point, 7, 9, 23

37
 © firstperson, inc.

performing a time-consuming job that the program doesn’t need to complete

before going on. When a thread has finished its job, the thread is suspended or
destroyed. See also process.

Unicode
A 16-bit character set defined by ISO 10646.

variable
An item of data named by an identifier. Each variable has a type, such as int or
Object, and a scope. See also class variable, instance variable, local variable.

36 Confidential
 © firstperson, inc.

octal
The numbering system using 8 as its base, using the numerals 0-7 as its digits. In
Oak programs, octal numbers must be preceded with 0. See also hexadecimal.

overloading
Using one identifier to refer to multiple items in the same scope. In the Oak
language, you can overload methods but not variables or operators.

overriding
Providing a different implementation of a method in a subclass of the class that
originally defined the method.

package
In the Oak language, a group of classes. Packages are declared with the package
keyword.

pixel
The smallest addressable picture element on a display screen or printed page.

pointer
A data element whose value is an address.

process
A virtual address space containing one or more threads.

root
In a hierarchy of items, the one item from which all other items are descended.
The root item has nothing above it in the hierarchy. See also hierarchy, class,
package.

scope
A characteristic of an identifier that determines where the identifier can be used.
Most identifiers in the Oak language have either class or local scope. Instance and
class variables and methods have class scope; they can be used outside the class
and its subclasses only by prefixing them with an instance of the class or (for class
variables and methods) with the class name. All other variables are declared
within methods and have local scope; they can be used only within the enclosing
block.

single precision
In the Oak language specification, describes a floating point number with 32 bits
of data. See also double precision.

subclass
A class that is derived from a particular class, perhaps with one or more classes in
between. See also superclass.

superclass
A class from which a particular class is derived, perhaps with one or more classes
in between. See also subclass.

thread
The basic unit of program execution. A process can have several threads running
concurrently, each performing a different job, such as waiting for events or

Confidential 35
 © firstperson, inc.

instance method
Any method that can be invoked using an instance of a class, but not using the
class name. Instance methods are defined in class definitions. See also class
method.

instance variable
Any item of data that’s associated with a particular object. Each instance of a class
has its own copy of the instance variables defined in the class. See also class
variable.

interface
In the Oak language, a group of methods that can be implemented by several
classes, regardless of where the classes are in the class hierarchy.

interpreter
A module that alternately decodes and executes every statement in some body of
code. The Oak interpreter decodes and executes Oak bytecode. See also compiler,
runtime system.

lexical
Pertaining to how the characters in source code are translated into tokens that the
compiler can understand.

linker
A module that builds an executable, complete program from component machine
code modules. The Oak linker creates a runnable program from compiled classes.
See also compiler, interpreter, runtime system.

literal
The basic representation of any integer, floating point, or character value. For
example, 3.0 is a single-precision floating point literal, and ‘a’ is a character literal.

local variable
A data item known within a block, but inaccessible to code outside the block. For
example, any variable defined within an Oak method is a local variable and can’t
be used outside the method.

method
A function defined in a class. See also instance method, class method.

multithreaded
Describes a program that is designed to have parts of its code execute
concurrently. See also thread.

object
The principle building blocks of object-oriented programs. Each object is a
programming unit consisting of data (instance variables) and functionality
(instance methods). See also class.

object oriented design

A software design method that models the characteristics of abstract or real
objects using classes and objects.

34 Confidential
 © firstperson, inc.

double precision
In the Oak language specification, describes a floating point number that holds 64
bits of data. See also single precision.

encapsulation
The localization of knowledge within a module. Because objects encapsulate data
and implementation, the user of an object can view the object as a black box that
provides services. Instance variables and methods can be added, deleted, or
changed, but as long as the services provided by the object remain the same, code
that uses the object can continue to use it without being rewritten.

exception
An event during program execution that prevents the program from continuing
normally; generally, an error. The Oak language supports exceptions with the try,
catch, and throw keywords. See also exception handler.

exception handler
A block of code that reacts to a specific type of exception. If the exception is for an
error that the program can recover from, the program can resume executing after
the exception handler has executed. See also exception.

formal parameter list

The parameters specified in the definition of a particular method. See also actual
parameter list.

garbage collection
The automatic detection and freeing of memory that is no longer in use. The Oak
runtime system performs garbage collection so that programmers never explicitly
free objects and other data.

hexadecimal
The numbering system that uses 16 as its base. The marks 0-9 and a-f (or
equivalently A-F) represent the digits 0 through 15. In Oak programs,
hexadecimal numbers must be preceded with 0x. See also octal.

hierarchy
A classification of relationships in which each item except the top one (known as
the root) is a specialized form of the item above it. Each item can have one or more
items below it in the hierarchy. In the Oak class hierarchy, the root is the Object
class.

identifier
The name of an item in an Oak program.

inheritance
The concept of classes automatically containing the variables and methods
defined in their superclasses.

instance
An object of a particular class. In Oak programs, an instance of a class is created
using the new operator followed by the class name.

Confidential 33
 © firstperson, inc.

class
In the Oak language, a type that defines the implementation of a particular kind
of object. A class definition defines instance and class variables and methods, as
well as specifying the interfaces the class implements and the immediate
superclass of the class.

class method
Any method that can be invoked using the name of a particular class. Class
methods affect the class as a whole, not a particular instance of the class. Class
methods are defined in class definitions. See also instance method.

class variable
A data item associated with a particular class as a whole—not with particular
instances of the class. Class variables are defined in class definitions. See also
instance variable.

comment
In a program, explanatory text that is ignored by the compiler. In Oak programs,
comments are delimited using // or /*...*/.

compilation unit
The smallest unit of Oak code that can be compiled. In the current Oak
implementation, the compilation unit is a file.

compiler
A program to translate source code into code to be executed by a computer. The
Oak compiler translates Oak source code into Oak bytecode. See also interpreter.

constructor
A method that creates an object. In the Oak language, constructors are instance
methods with the same name as their class. Oak constructors are invoked using
the new keyword.

critical section
A segment of code in which a thread uses resources (such as certain instance
variables) that can be used by other threads, but that must not be used by them at
the same time.

declaration
A statement that establishes an identifier and associates attributes with it, without
necessarily reserving its storage (for data) or providing the implementation (for
methods). See also definition.

definition
A declaration that reserves storage (for data) or provides implementation (for
methods).

derived from
Describes a class that inherits properties of another class. See also subclass,
superclass.

distributed
Running in more than one address space.

32 Confidential
 © firstperson, inc.

Glossary
abstract class
A class that should never be instantiated; only its subclasses should be
instantiated. Abstract classes are defined so that other classes can inherit from
them.

actual parameter list

The arguments specified in a particular method call. See also formal parameter list.

argument
A data item specified in a method call. An argument can be a literal value, a
variable, or an expression.

array
A collection of data items, all of the same type, in which each item’s position is
uniquely designated by an integer.

ASCII
American Standard Code for Information Interchange. A standard assignment of
7-bit numeric codes to characters. See also Unicode.

atomic
Refers to an operation that is never interrupted or left in an incomplete state
under any circumstance.

binary operator
An operator that has two arguments.

bit
The smallest unit of information in a computer, with a value of either 0 or 1.

bitwise operator
An operator that manipulates bit-oriented data, such as by performing the logical
AND operation such that each bit that’s 1 in either operand is 1 in the result.

block
In the Oak language, any code between matching braces ({ and }).

boolean
Refers to an expression or variable that can have only a true or false value. The
Oak language provides the boolean type and the literal values true and false.

byte
A sequence of eight bits. The Oak language provides a corresponding byte type.

bytecode
Machine-independent code generated by the Oak compiler and executed by the
Oak interpreter.

casting
Explicit conversion from one data type to another.

Confidential 31
 A Appendix: Floating Point © firstperson, inc.

A.3 Ordering

The usual relational operators can be applied to floating-point values. With the
exception of NaN, all floating values are ordered, with -Inf < all finite values <
Inf.

-Inf == -Inf, +Inf == +Inf, -0. == 0. The ordering relations are transitive.
Equality and inequality are reflexive.

NaN is unordered. Thus the result of any order relation between NaN and any
other value is false and produces 0. The one exception is that “NaN != anything”
is true.

Note that, because NaN is unordered, Oak’s logical inversion operator, !, does not
distribute over floating point relationals as it can over integers.

A.4 Summary of IEEE-754 Differences

Oak arithmetic is a subset of the IEEE-754 standard. Here is a summary of the key
differences.
• Nonstop Arithmetic—The Oak system will not throw exceptions, traps, or
otherwise signal the IEEE exceptional conditions: invalid operation,
division by zero, overflow, underflow, or inexact. Oak has no signaling
NaN.
• Rounding—Oak rounds inexact results to the nearest representable value,
with ties going to the value with a 0 least-significant bit. This is the IEEE
default mode. But, Oak rounds towards zero when converting a floating
value to an integer. Oak does not provide the user-selectable rounding
modes for floating-point computations: up, down, or towards zero.
• Relational set—Oak has no relational predicates which include the
unordered condition, except for !=. However, all cases but one can be
constructed by the programmer, using the existing relations and logical
inversion. The exception case is ordered but unequal. There is no specific
IEEE requirement here.
• Extended formats—Oak does not support any extended formats, except
that double will serve as single-extended. Other extended formats are not a
requirement of the standard.

30 Confidential
 A Appendix: Floating Point © firstperson, inc.

A Appendix: Floating Point

This appendix discusses properties of Oak floating point arithmetic: general
precision notes and special values, binary format conversion, ordering. At the end
is a section summarizing the differences between Oak arithmetic and the IEEE 754
standard. For more information on the IEEE 754 standard, see “IEEE Standard for
Binary Floating-Point Arithmetic, ANSI/IEEE Std. 754-1985.”

Operations involving only single-precision float and integer values are

performed using at least single-precision arithmetic and produce a single-
precision result. Other operations are performed in double precision and produce
a double precision result. Oak floating-point arithmetic produces no exceptions.

Underflow is gradual.

A.1 Special Values

There is both a positive zero and a negative zero. The latter can be produced in a
number of special circumstances: the total underflow of a * or / of terms of
different sign; the addition of -0 to itself or subtraction of positive zero from it; the
square root of -0. Converting -0 to a string results in a leading ‘-’. Apart from this,
the two zeros are indistinguishable.

Calculations which would produce a value beyond the range of the arithmetic
being used deliver a signed infinite result. An infinity (Inf) has a larger
magnitude than any value with the same sign. Infinities of the same sign cannot
be distinguished. Thus, for instance (1./0.) + (1./0.) == (1./0.). Division
of a finite value by infinity yields a 0 result.

Calculations which cannot produce any meaningful numeric result deliver a

distinguished result called Not A Number (NaN). Any operation having a NaN as
an operand produces a NaN as the result. NaN is not signed and not ordered (see
“Ordering” on page 30). Division of infinity by infinity yields NaN, as does
subtraction of one infinity from another of the same sign.

A.2 Binary Format Conversion

Converting a floating-point value to an integer format results in a value with the

same sign as the argument value and having the largest magnitude less than or
equal to that of the argument. In other words, conversion rounds towards zero.
Converting infinity or any value beyond the range of the target integer type gives
a result having the same sign as the argument and the maximum magnitude of
that sign. Converting NaN results in 0.

Converting an integer to a floating format results in the closest possible value in

the target format. Ties are broken in favor of the most even value (having 0 as the
least-significant bit).

Confidential 29
 10 Garbage Collection © firstperson, inc.

The finally statement is executed even if the try block contains a goto, return,
break, continue, or throw statement. For example, the following code example
always results in “finally” being printed, but “after try” is never printed.

try {
if (a==10)
return;
} finally {
print("finally\n");
}
print("after try\n");

9.4.2 Asynchronous Exceptions

Generally, exceptions are synchronous—they are thrown by code executed

sequentially by an Oak program. However, in programs that have multiple
threads of execution, one thread can throw an exception (using Thread’s
postException() instance method) to another thread. The second thread can’t
predict exactly when it will be thrown an exception, so the exception is
asynchronous.

Implementation Note: As of Oak 0.2, no FirstPerson-supplied code throws

asynchronous exceptions, so you don’t need to worry about them unless you use
them in your own code.

The default will probably be By default, asynchronous exceptions can happen at any time. To prevent
changed to not allow asynchro- asynchronous exceptions from occuring in a critical section of code, you can mark
nous exceptions except in ex-
plicitly unprotected sections of the code with the protect keyword, as shown below:
code.
protect {
/* critical section goes here */
}

To allow asynchronous exceptions to occur in an otherwise protected section of

code, use the unprotect keyword, as follows:

unprotect {
/* code that can afford asynchronous exceptions */
}

10 Garbage Collection
The Oak garbage collector makes most aspects of storage management simple
and robust. Oak programs never need to explicitly free storage: it is done for them
automatically. The garbage collector never frees pieces of memory that are still
referenced, and it always frees pieces that are not. This makes both dangling
pointer bugs and storage leaks impossible. It also frees designers from having to
figure out which parts of a system have to be responsible for managing storage.

The garbage collector also does compaction: it copies all objects to the beginning
of the heap, coalescing free space in one large chunk at the end. This eliminates
the loss of free space due to fragmentation.

28 Oak Language Specification Confidential

 9 Statements © firstperson, inc.

for an exception handler to resume execution at the point that the exception
occurred.

For example, this code fragment:

class Foo {};

print("now ");
try {
print("is ");
throw new Foo();
print("a ");
} catch(Foo p) {
print("the ");
}
print("time\n");

prints “now is the time”. As this example shows, exceptions don’t have to be used
only for error handling, but any other use is likely to result in code that’s hard to
understand.

Exception handlers can be nested, allowing exception handling to happen in more

than one place. Nested exception handling is often used when the first handler
can’t recover completely from the error, and yet needs to execute some cleanup
code (as shown in the following code example). To pass exception handling up to
the next higher handler, use the throw keyword without specifying a
GenericException instance. Note that the method that rethrows the exception
stops executing after the throw statement; it never returns.

try {
f.open();
} catch(GenericException e) {
f.close();
throw;
}

9.4.1 The finally Statement

The following example shows the use of a finally statement that is useful for
guaranteeing that some code gets executed whether or not an exception occurs.
You can use either a catch statement or a finally statement within a particular try
block, but not both. For example, the following code example:

try {
/* do something */
} finally {
/* clean up after it */
}

is similar to:

try {
/* do something */
} catch(Object e){
/* clean up after it */
throw;
}
/* clean up after it */

Confidential Oak Language Specification 27

 9 Statements © firstperson, inc.

continue;

return e1;

for(e1; e2; e3) statement

while(boolean) statement

do statement
while(boolean);

9.4 Exceptions1

When an error occurs in an Oak program—for example, when an argument has

an invalid value—the code that detects the error can throw an exception. By
default, exceptions result in the thread terminating with an error message.
However, programs can have exception handlers that catch the exception and
recover from the error.

Some exceptions are thrown by the Oak runtime system. However, any class can
define its own exceptions and cause them to occur using throw statements. A
throw statement consists of the throw keyword followed by an object. By
convention, the object should be an instance of GenericException or one of its
subclasses. The throw statement causes execution to switch to the appropriate
exception handler. When a throw statement is executed, any code following it is
not executed, and no value is returned by its enclosing method. The following
example shows how to create a subclass of GenericException and throw an
exception.

class MyException extends GenericException {};

if (/* no error occurred */)

/* do something */
else /* error occurred */
throw new MyException();

To define an exception handler, the program must first surround the code that can
cause the exception with a try statement. After the try statement come one or
more catch clauses—one per exception class that the program can handle at that
point. In each catch clause is exception handling code. For example:

try {
p.a = 10;
} catch (NullPointerException e) {
print("p was null\n");
} catch (GenericException e) {
print("other error occurred\n");
}

A catch clause is like a method definition with exactly one parameter and no
return type. When an exception occurs, the runtime system searches the nested
try/catch clauses. The first one with a parameter type that is the same class or a
superclass of the thrown object has its catch clause executed. After the catch
clause executes, execution resumes after the try/catch statement. It is not possible

1. Oak exception handling closely follows the proposal in the second edition of The C++ Programming
Language, by Bjarne Stroustrup.

26 Oak Language Specification Confidential

 9 Statements © firstperson, inc.

8.2 Casts and Conversions

The Oak language and runtime system restrict casts and conversions to help
prevent the possibility of corrupting the system. Integers and floating point
numbers can be cast back and forth, but integers cannot be cast to arrays or
objects. An instance can be cast to a superclass with no penalty, but casting to a
subclass generates a runtime check. If the object being cast to a subclass is not an
instance of the subclass (or one of its subclasses), the runtime system throws an
InvalidClassCastException.

9 Statements

9.1 Declarations

Declarations can appear anywhere that a statement is allowed. The scope of the
declaration ends at the end of the enclosing block.

In addition, declarations are allowed at the head of for statements, as shown

below:

for (int i = 0; i<10; i++) . . .

Items declared in this way are valid only within the scope of the for statement.
For example, the preceding code sample is equivalent to the following:

{
int i = 0;
for (; i<10; i++) . . .
}

9.2 Expressions

As in C, expressions are statements:

a = 3;
print(23);

9.3 Control Flow

Except for the for statement, which can contain declarations (as described in
“Declarations” on page 25), this is just like C:

if(boolean) statement
else statement

switch(e1) {
case e2: statements
default: statements
}

break;

goto label;

Confidential Oak Language Specification 25

 8 Expressions © firstperson, inc.

results. Floating point operations cannot cause exceptions, but they can produce
the special values of Infinity or Not-a-Number.

The usual relational operators are also available, and produce boolean results: >,
<, >=, <=, ==, !=. Because of the properties of Not-a-Number, floating point values
are not fully ordered, so care must be taken in comparison. For instance, if a<b is
not true, it does not follow that a>=b. Likewise, a!=b does not imply that a>b ||
a<b. In fact, there may no ordering at all.

Floating point arithmetic and data formats are defined by IEEE 754, “Standard for
Floating Point Arithmetic.” See “Appendix: Floating Point” on page 29 for details
on the Oak language’s floating point implementation.

8.1.4 Operators on Strings

The operator + concatenates Strings, automatically converting operands into

Strings if necessary.

float a = 1.0;
print("The value of a is " + a + "\n");
print("" + 1.01 + 2 + "\n");
print(1.01 + 2 + "\n"); // = (1.01 + 2) + "\n"

The value of a is 1
1.012
3.01

The += and ++ operators also work on Strings. They and their rough equivalents
are shown below. Note, however, that the left hand side (s1 in the following
examples) is evaluated only once.

s1 += a; //s1 = s1 + a; a is converted to String if necessary

s1++; //s1 = s1 + "1"

print() and println() aren’t re- The print() operator prints the argument specified in its parentheses, coercing it to
ally operators; they’re Oak-de-
fined global methods. Such be a String if necessary. The println() operator is the same as print(), except that it
global methods are expected to adds "\n" (the newline character) to the end of the specified String.
be replaced by class-based
methods.
8.1.5 Operators on Objects

The unary operator clone is applied to an object. It atomically allocates space for a
new instance of the same class and copies the contents of the existing object into
it, making the new object an exact, shallow copy of the old one. For example, if the
existing object refers to another object, the clone refers to the same object—the
referred-to object is not cloned. The clone operator is normally used inside new to
clone the prototype of some class, before applying the initializers (constructors).

instanceof might be replaced The binary operator instanceof tests whether the specified object is an instance of
by an operator that is more dy- the specified class or one of its subclasses. For example,
namic. It might become a meth-
od.
(thermostat instanceof MeasuringDevice)

determines whether thermostat is a MeasuringDevice object (an instance of

MeasuringDevice or one of its subclasses).

24 Oak Language Specification Confidential

 8 Expressions © firstperson, inc.

Table 2. Binary Integer Operators: integer op integer ⇒ integer

Operator Operation
+ addition
– subtraction
* multiplication
/ division
% modulus
& bitwise AND
| bitwise OR
^ bitwise XOR
<< left shift
sign-propagating
>>
right shift
>>> zero-fill right shift

Integer division rounds toward zero. Division and modulus obey the identity
(a/b)*b + (a%b) == a. Although it may not be obvious that % could overflow,
it does for a zero divisor.

An op= assignment operator corresponds to each of the binary operators in the

above table.

The integer relational operators <, >, <=, >=, ==, and != produce boolean results.
They cannot overflow.

8.1.2 Operators on Boolean Values

Variables or expressions that are boolean can be combined to yield other boolean
The &, |, and ^ operators values. The unary operator ! is boolean negation. The binary operators &, |, and ^
aren’t implemented yet for are the logical AND, OR, and XOR operators; they force evaluation of both
boolean values..
operands. To avoid evaluation of right-hand operands, you can use the short-cut
evaluation operators && and ||. You can also use == and !=. The assignment
operators also work: &=, |=, ^=. The conditional operator ?: works much as it
does in C.

8.1.3 Operators on Floating Point Values

Double precision and spe- Floating point values can be combined using the usual operators: unary –; binary
cial mathematical values +, –, *, and /; and the assignment operators +=, –=, *=, and /=. The ++ and --
are not implemented yet.
operators also work on floating point values (they add or subtract 1.0). In
addition, % and %= work on floating point values. Operators that work on
integers but that aren’t listed in this section work on floating point values by first
converting the floating point values into integers.

Floating point expressions involving only single-precision operands are

evaluated using single-precision operations and produce single-precision results.
Floating point expressions that involve at least one double-precision operand are
evaluated using double-precision operations and produce double-precision

Confidential Oak Language Specification 23

 8 Expressions © firstperson, inc.

8 Expressions
Expressions in the Oak language are much like expressions in C.

8.1 Operators

The Oak operators, from highest to lowest priority, are:

. [] ()
++ -- ! ~ instanceof new clone
* / %
+ -
<< >> >>>
< > <= >=
== !=
&
^
|
&&
||
?:
= op=
,

8.1.1 Operators on Integers

For operators with integer results, if any operand is long, the result type is long.
Otherwise the result type is int—never byte or short. When a result outside an
operator’s range would be produced, the result is reduced modulo the range of
the result type.

Table 1. Unary Integer Operators: op integer ⇒ integer

Operator Operation
– unary negation
~ bitwise complement

22 Oak Language Specification Confidential

 7 Assertions © firstperson, inc.

7 Assertions
Assertions aren’t imple- The Oak language has a set of facilities that allow assertions to be made about the
mented yet. behavior of programs. These allow extensive checking and a corresponding
increase in the reliability of programs. A failed assertion results in an
AssertionFailedException (see “Exceptions” on page 26).

7.1 Constraints on Instance Variables and Methods

The assert keyword can be used to declare a set of constraints on instance
variables and methods. This enables concise documentation of a class designer’s
intentions. The annotations also serve as a binding contract between a class
designer and a class maintainer.

While objects are not required to obey the legality constraints within methods, the
constraints are enforced at the entry and exit of every public and protected
method. In classes that use assert for all variables, all public and protected
methods can expect to operate on a coherent object and have the responsibility of
restoring coherence before finishing. The following example shows how to use
assert to constrain the values of two instance variables.

class Calender {
static int lastDay[12]=
{31,29,31,30,31,30,31,31,30,31,30,31};
int month assert(month >=1 && month <=12);
int date assert(date>=1 && date<=lastDay[month]);
}

7.2 Preconditions and Postconditions

The syntax described in The behavior of a method can be specified by a set of preconditions that must
this section might change hold before the method begins and a set of postconditions that must hold after it
when preconditions and
postconditions are imple- finishes.
mented.
class Stack {
int length;
Element element[];
boolean full() { /* . . . */ };
boolean empty() { return length==0; }

Element pop() {
precondition: !empty();
/* . . . */
postcondition: !full();
}

void push(Element x) {
precondition: !full();
/* . . . */
postcondition: !empty();
}
}

Preconditions and postconditions are inherited by subclasses: methods

overridden by a subclass must obey the preconditions and postconditions of their
superclass. Inherited preconditions and postconditions cannot be restricted or
redefined.

Confidential Oak Language Specification 21

 6 Packages © firstperson, inc.

Packages” on page 20.) The oak.lang package is documented in the FirstPerson

Programming Interface in the chapter “The Language (oak.lang) Package.”

Implementation Note: In the current implementation of Oak, packages are

closely tied to the file system. Everything that’s in a particular package is in one
directory, and only one package can be in a particular directory. The name of a
package indicates its directory. For example, a package named oak.lang would
be in a directory lang under a directory named oak . The oak/lang directory
could be anywhere in the file system. The Oak runtime system looks for it in the
directories specified by the CLASSPATH environment variable.

6.1 Specifying a Compilation Unit’s Package

The package that a compilation unit is in is specified by a package statement. It

has the following format:

package packageName;

When a compilation unit has no package statement, the unit is placed in a default
package, which has no name.

A compilation unit automatically imports every class and interface in its own
package.

6.2 Using Classes and Interfaces from Other Packages

Code in one package can specify classes or interfaces from another package in one
of two ways:
• By prefacing each reference to the class or interface name with the name of
its package

// prefacing with a package

myCo.aGroup.AClass o=new myCo.aGroup.AClass();

• By importing the class or interface or the package that contains it, using an
import statement. Importing a class or interface makes the name of the class
or interface available in the current namespace. Importing a package makes
the names of all of its public classes and interfaces available.

// importing just a class

import myCo.aGroup.AClass;
AClass o=new AClass();

The syntax for importing // importing an entire package

packages is likely to import myCo.aGroup.*;
change. AClass o=new AClass();

20 Oak Language Specification Confidential

 6 Packages © firstperson, inc.

the programmer specify that an object must implement a given interface, without
having to know the exact type or inheritance of that object. Using interfaces
makes it unnecessary to force related classes to share a common abstract
superclass or to add methods to Object just to guarantee that many classes
implement the same methods.

class StorageManager {
Stream stream;
...
void pickle(Storing obj) {
obj.freezeDry(stream);
}
}

5.2 Constants in Interfaces

Besides methods, interfaces can also declare constants. The value of the constant
must be set in the interface. For example:

interface InterfaceName {
const int aConstant = 42;
. . .
}

Code can refer to the interface constants as if they were declared as class (static)
variables in the implementing class. When Oak detects a class implementing two
or more interfaces that declare constants with the same name, an error results.
One way to avoid the possibility of this error is to specify the interface name
before every use of the constant, whether in the implementing class or in clients:

InterfaceName.aConstant
o.aConstant // where o is declared as "InterfaceName o;"

5.3 Combining Interfaces

Interfaces can incorporate one or more other interfaces, using the extends
keyword as follows:

interface DoesItAll extends Storing, Painting {

void doesSomethingElse() = 0;
}

6 Packages
Packages are groups of classes and interfaces. They are a tool for managing a large
namespace and avoiding conflicts. Every class and interface name is contained in
some package. (See “Classes” on page 10 and “Access to Variables and Methods”
on page 15 for information on how packages affect the namespace.) By
convention, package names consist of period-separated words, with the first
name representing the organization that developed the package.

The oak.lang package contains classes and interfaces that are integral to the Oak
language. All Oak programs automatically import the contents of the oak.lang
package. (Importing is discussed in “Using Classes and Interfaces from Other

Confidential Oak Language Specification 19

 5 Interfaces © firstperson, inc.

class A {
void a() { f.set(42); }
B f;
}

class B {
void set(long n) { N = n; }
long N;
}

5 Interfaces
An interface specifies a collection of methods without implementing their bodies.
Interfaces provide encapsulation of method protocols without restricting the
implementation to one inheritance tree. When a class implements an interface, it
generally must implement the bodies of all the methods described in the interface.
(The exception is that if the implementing class is abstract—never instantiated—it
can leave the implementation of some or all interface methods to its subclasses.)

Interfaces solve some of the same problems that multiple inheritance does
without as much overhead at runtime. However, because interfaces involve
dynamic method binding, there is often a small performance penalty to using
them.

Using interfaces allows several classes to share a programming interface without

having to be fully aware of each other’s implementation. The following example
shows an interface declaration (with the interface keyword) and a class that
implements the interface.

In the future, the “=0” public interface Storing {

part of declaring meth- void freezeDry(Stream s) = 0;
ods in interfaces may go void reconstitute(Stream s) = 0;
away. }

public class Raster implements Storing, Painting {

...
void freezeDry(Stream s) {
/* JPEG compress image before storing */
...
}

void reconstitute (stream s) {

/* JPEG decompress image before reading */
...
}
}

Like classes, interfaces are either private (the default) or public. The scope of
public and private interfaces is the same as that of public and private classes,
respectively. As for classes, the public and private keywords specify whether an
interface is public or private.

5.1 Interfaces as Types

The declaration syntax interfaceName variableName declares a variable or

parameter to be an instance of some class that implements interfaceName. This lets

18 Oak Language Specification Confidential

 4 Classes © firstperson, inc.

class MyClass {
. . .
MyClass(someParameters) {
/* Call immediate superclass constructor */
super(otherParameters);
. . .
}
. . .
}

Invoking a constructor in the current class is done as shown in the following code
example.

class MyClass {
. . .
MyClass(someParameters) {
. . .
}
MyClass(otherParameters) {
/* Call the constructor in this class that has the
specified parameter list. */
this(someParameters);
. . .
}
. . .
}

The Foo and FooSub methods below are examples of constructors.

class Foo extends Bar {

int a;
Foo(int anInt) {
// implicit call to Bar()
a = anInt;
}
Foo() {
this(42); // calls Foo(42) instead of Bar()
}
}

class FooSub extends Foo {

int b;
FooSub(int anInt) {
super(13); // calls Foo(13); without this line,
// would have called Foo()
b = anInt;
}
}

If a class declares no constructors, the compiler automatically generates one of the

following form:

ClassName() {
super();
}

4.13 Order of Declarations

The order of declaration of classes and the methods and instance variables within
them is irrelevant. Methods are free to make forward references to other methods
and instance variables. The following works:

Confidential Oak Language Specification 17

 4 Classes © firstperson, inc.

synchronized [other access specifiers]* <return type>

<method name> (<parameter list>) {
/* implementation */
}
Currently, you must use
synchronize (no “d”) in- Synchronized blocks are declared as follows:
stead of synchronized
when declaring synchro- /* ...preceding code in the method... */
nized blocks. Synchro- synchronize(<object or class name>) { //sync. block
nized blocks based on /* code that requires synchronized access */
classes (as opposed to ob- }
jects) currently don’t work. /* ...remaining code in the method... */

When a synchronized method is invoked, it waits until it can acquire the lock for
the current instance (or class, if it’s a class method). After acquiring the lock, it
executes its code and then releases the lock.

Synchronized blocks of code behave similarly, except that instead of using the
lock for the current instance or class, they use the lock associated with the object
or class specified in the block’s synchronize statement.

For more information on blocks of code running simultaneously, see the Thread
class documentation in the FirstPerson Programming Interface.

4.12 Constructors

Constructors are special methods provided for initialization. They are

distinguished by having the same name as their class. Constructors are
automatically called upon the creation of an object. They cannot be called
explicitly through an object. Constructors do not have any return type.

Constructors can be overloaded by varying the number and types of parameters,

just as any other method can be overloaded.

Class Foo {
int x;
float y;
Foo() { x=0; y=0.0; }
Foo(int a) { x=a; y=0.0; }
Foo(float a) { x=0; y=a; }
Foo(int a, float b) { x=a; y=b; }
}

Foo obj1 = new Foo(); //calls Foo();

Foo obj2 = new Foo(4); //calls Foo(int a);
Foo obj3 = new Foo(4.0); //calls Foo(float a);
Foo obj4 = new Foo(4, 4.0); //calls Foo(int a, float b);

Before the constructor is called, storage for an instance is atomically allocated and
initialized to be a copy of the prototype for the class.

The instance variables of superclasses are initialized by calling either a

constructor for the immediate superclass or a constructor for the current class. If
neither is specified in the code, the superclass constructor that has no parameters
is invoked. Calling a constructor must be the first thing in the method body;
calling a constructor later is illegal.

Invoking a constructor of the immediate superclass is done as follows:

16 Oak Language Specification Confidential

 4 Classes © firstperson, inc.

4.8 Transient Variables

Variables marked transient are treated specially when instances of the class are
written out as persistent objects. Specifically, the values of transient variables are
not written out. When the persistent object is reconstituted, transient variables are
initialized to zero.

4.9 Final Classes and Methods

The final keyword is an access specifier that marks a class as never having
subclasses, or a method as never being overridden. Using final lets the compiler
perform a variety of optimizations. One such optimization is inline expansion of
method bodies, which is done for small, final methods (where the meaning of
small is implementation dependent).

4.10 Access to Variables and Methods

Each variable or method declared in a class has one of the following types of
access: public, protected, or private. These access types affect whether the
variable or method can be used by other classes.

The rules for access to Note: All classes in a particular package can use all variables and methods
classes and interfaces in declared in the classes in that package, regardless of public, protected, and
the same package might
change. private declarations (see “Packages” on page 19).

By default all variables and methods in a class (including constructors) are

private. Private variables and methods can be accessed only by methods declared
in the class, and not by its subclasses or any other classes (except for classes in the
same package). Public variables and methods—those declared with the public
type modifier—can be accessed by anyone. The protected type modifier makes a
variable or method accessible to subclasses, but not to any other classes (except
those in the same package).

The following example shows how to specify access.

class Stuff {
int i; /* private by default */
public int j; /* visible to everyone */
protected int k; /* subclasses see this */
void method1() { } /* private by default */
public void method2() { } /* public */
protected static void method3() { } /* protected */
};

4.11 Synchronized Methods and Blocks

The synchronized keyword is an access specifier that marks a method or block of

code as being required to acquire a lock, so that it does not run at the same time as
other code that needs access to the same resource. (The other code must also be
marked synchronized.) Each object has exactly one lock associated with it; each
class also has exactly one lock.

Synchronized methods are declared as follows:

Confidential Oak Language Specification 15

 4 Classes © firstperson, inc.

class Ahem {
int i; // Instance variable
static int j; // Class variable
void seti(int I) { i = I; } // Instance method
static void setj(int J) { j = J; } // Class method
};

Ahem a = new Ahem();

Ahem.j = 2; /* valid; class var via class */
a.j = 3; /* valid; class var via instance */
Ahem.setj(2); /* valid; class method via class */
a.setj(3); /* valid; class method via instance */
a.i = 4; /* valid; instance var via instance */
Ahem.i = 5; /* ERROR; instance var via class */
a.seti(4); /* valid; instance method via instance */
Ahem.seti(5); /* ERROR; instance method via class */

A class variable exists only once per address space, no matter how many
instances of the class exist in that address space. For distributed applications that
run in multiple address spaces, each address space has one occurrence of the class
variable. When you refer to a class variable relative to some object (for example,
obj.aVar) the class variable aVar is fetched from the address space where obj
resides. See “Synchronized Methods and Blocks” on page 15 for information on
achieving synchronized access to class variables.

Class variables can have initializers, just as instance variables can. These
initializers are executed just before the first runtime use of the class, before any
instances are created. You can also add a code fragment to be executed at the same
time the class variables are initialized, as shown in the following example.

class A {
static int arr[12];
static { /* code fragment: initialize the array */
int i;
for (i = 0; i<arr.length; i++)
arr[i] = i;
}
}

Class methods cannot refer to instance variables; they can only use class variables.

4.6 Constants

Instance and class variables can be marked const to indicate that, once initialized,
their value never changes.

4.7 Volatile Variables

Instance and class variables can be marked volatile so that the compiler treats
them specially during optimization. The values of volatile variables are never
cached in registers and are always re-read when referenced. Variables should be
marked volatile when they might be changed by means undetectable by the
compiler, such as by another thread or device.

14 Oak Language Specification Confidential

 4 Classes © firstperson, inc.

4.3 Overriding and Overloading Methods

The Oak language allows polymorphic method naming—declaring a method with

a name that has already been used in the class or its superclass—for overriding
and overloading methods. Overriding means providing a different
implementation of an inherited method. Overloading means declaring a method
that has the same name as another method, but a different parameter list.

Note: Return types are not used to distinguish methods. Within a class scope,
methods that have the same name and parameter list must return the same type.

To override a method, a subclass of the class that originally declared the method
must declare a method with the same name, return type (or a subclass), and
parameter list. When the method is invoked on an instance of the subclass, the
new method is called rather than the original method.

To overload a method, a class declares a method that has the same name and
return type as another method (which has been declared in the class or in one of
its superclasses), but a different parameter list. The Oak runtime system resolves
which method to call by matching the actual parameter list (the parameter list
passed to the method) against the formal parameter lists of all methods with the
same name.

class A {
void Thermostat(Foo f) {}
}
class B extends A {
void Thermostat(Foo f) {} // override
void Thermostat() {} // overload
int Thermostat() {} // ERROR: Duplicate method
}

When deciding which method to invoke, the runtime system computes the
number of conversions required to change the actual parameter list into the types
declared in each method’s formal parameter list. The method that requires the
fewest conversions is chosen. If there is a tie, the method call is ambiguous and a
compilation error occurs.

Note: The names of parameters are not significant. Only the number, type, and
order are.

4.4 Used before Set

Methods are rigorously checked to be sure that all local variables (variables
declared inside a method) are set before they are referenced. Used-before-set is a
fatal compilation error.

4.5 Class Variables and Methods

Variables and methods declared in a class can be declared static, which makes
them apply to the class itself, rather than to an instance of the class. As shown in
the following code example, both class variables and class methods are accessed
using the class name. For convenience, they can also be accessed using an instance
of the class.

Confidential Oak Language Specification 13

 4 Classes © firstperson, inc.

void aMethod() {
/* . . . */
otherObject.Method(this);
/* . . . */
}

Any time a method refers to its own instance variables or methods an implicit
“this.” is in front of each reference:

class Foo {
int a, b, c;
/* . . . */
print(a + "\n"); // a == "this.a"
/* . . . */
}

Instance variables can’t be hidden by being redeclared in subclasses. Specifically,

if a class declares a public or protected instance variable, that variable cannot be
redeclared in any subclass, although it can be used by any subclass. (See “Access
to Variables and Methods” on page 15 for information on declaring variables
public, protected, and private.) Private instance variables can be redeclared, since
they aren’t visible to subclasses.

4.2 Methods

Methods are the operations that can be performed on an object or class. They can
be declared in either classes or interfaces, but they can be implemented only in
classes. (Note: All user-defined operations in Oak are implemented with methods;
Oak has no functions.)

A method declaration in a class has the following form:

[accessSpecifiers] returnType methodName ( parameterList ) {

[methodBody]
}

A method declaration in an interface has the following form:

[accessSpecifiers] returnType methodName ( parameterList ) = 0;

Methods:
• Have a return type unless they’re constructors, in which case they must
have no return type. If a non-constructor method does not return any value,
it must have a void return type.
• Have a parameter list consisting of comma-separated pairs of types and
parameters. The parameter list should be empty if the method has no
parameters.
In the Oak 0.2 release, local
variables and parameters can’t Variables declared in methods (local variables) can’t hide other local variables or
hide instance variable names. parameters. For example, if a method is implemented with a parameter named i,
In the future, this restriction it’s a compile-time error for the method to declare a variable named i.
will be eased; the compiler will
generate a warning instead of
an error.

12 Oak Language Specification Confidential

 4 Classes © firstperson, inc.

conversion or cast. Casts from a class to a subclass are always checked to make
sure that the object is actually an instance of the subclass (or one of its subclasses).

The Oak language supports single inheritance. Through a feature known as

interfaces, it supports some features that in other languages are supported through
multiple inheritance (see “Interfaces” on page 18). Instances of classes are stored
in a garbage collected heap (see “Garbage Collection” on page 28); local variables
are references to objects in the heap.

The immediate superclass of a class and the interfaces that the class implements
(if any) are indicated in the class declaration by the keywords extends and
implements, respectively:

public class Classname extends Superclassname

implements Interface1, Interface2 {
/* . . . */
}

Every class except the root class has exactly one immediate superclass. Unlike in
C++, all Oak classes are derived from a single root class: Object. If a class is
declared without specifying an immediate superclass, Object is assumed. For
example, the following

public class Point {

float x, y;
}

is the same as

public class Point extends Object {

float x, y;
}

Classes are either private (the default) or public. Private classes are invisible
outside of the package in which they’re declared. Public classes can be used
outside of their package. Public classes can’t be derived from private classes. To
declare a class public (or private), use the public (or private) keyword.

4.1 Instance Variables

Instance variables are declared just like local variables. They can be of any type
and can have initializers. These initializers are executed when the instance is
initialized. (If an instance variable does not have an initializer, it is initialized to
zero or, for boolean variables, to false.) An example of an initializer for an
instance variable named j follows.

class A {
int j = 23;
/* . . . */
}
Inside the scope of an instance of a class, the name this represents the current
object. For example, an object may need to pass itself as an argument to another
object’s method:

Confidential Oak Language Specification 11

 4 Classes © firstperson, inc.

3.4 Character Types

Unicode is not implement- The Oak language uses the Unicode character set throughout. Consequently the
ed yet. Characters current- char data type is defined as a 16-bit unsigned integer.
ly have ASCII values,
although they are stored in
16 bits.
3.5 Arrays

The Oak language includes support for arrays—sets of ordered data items. Arrays
are referred to and passed by reference.

Subscripts are checked to make sure they’re valid:

int a[10];
a[5] = 1;
a[11] = 2;/* ERROR */

Array dimensions can be integer expressions:

void doIt(int n) {
float arr[n];
...
}

The length of any array can be found by using .length:

int a[10][3];
The print operator is one print(a.length + ", " + a[0].length + "\n");
of a group of operators
whose functionality might 10, 3
be moved into classes.
Arrays are allocated either where they’re declared (by specifying the dimensions
of the array when it is declared, as shown above) or dynamically with the new
keyword:

Although arrays can be int a[];

created with new, just as a = new int[10];
instances of classes are cre- Raster foo[];
ated, arrays are not cur- foo = new Raster[10]; //creates an array, but not the
rently objects. //Raster objects in the array.
foo[1] = new Raster("blah.jpg");

4 Classes
Classes represent the classical object oriented programming model. They support
data abstraction and implementations tied to data.

To make a new class, the programmer must base it on an existing class. The new
class is said to be derived from the existing class. The derived class is also called a
subclass of the other, which is known as a superclass. Class derivation is transitive:
if B is a subclass of A, and C is a subclass of B, then C is a subclass of A.

If B is a subclass of A, then an instance of B can be used as an instance of A. In fact,

if there is no ambiguity, then no explicit cast is needed. If an instance of A needs
to be used as if it were an instance of B, the programmer can write a type

10 Oak Language Specification Confidential

 3 Types © firstperson, inc.

introduced. The four integer types have widths of 8, 16, 32, and 64 bits, and are
unsigned isn’t implement- signed unless prefixed by the unsigned modifier.
ed yet; it might never be.
Width Name Comments
The Oak byte type is what C programmers are used
to thinking of as the char type. But in the Oak lan-
guage, characters are 16 bits wide. Having a sepa-
8 byte
rate byte type removes the confusion in C between
the interpretation of char as an 8 bit integer and as
a character.
In C, the width of short is generally 16 bits, but the
16 short C specification says it can be larger. In Oak, short is
always 16 bits wide.
An int in the Oak language is always 32 bits wide.
In C, the width of int is implementation defined
32 int
and is most often 32 bits, but is sometimes 16 bits,
and has been other values (such as 60).
The Oak language’s definition of long is a break
from the C tradition that specifies that long is 32
bits and long long is 64 bits. With the standardiza-
64 long
tion of int to mean 32 bits, it is redundant to have
two types with the same meaning and unnecessary
to have such an odd type name for 64 bits.

A variable’s type does not directly affect its storage allocation. Type only
determines the variable’s arithmetic properties and legal range of values. If a
value is assigned to a variable that is outside the legal range of the variable, the
Value reduction is not im- value is reduced modulo the range.
plemented yet.

3.2 Floating Point Types

double is not implemented The float keyword denotes single precision (32 bit); double denotes double
yet. precision (64 bit). The result of a binary operator on two float operands is a float.
If either operand is a double, the result is a double.

Floating point arithmetic and data formats are defined by IEEE 754. See
“Appendix: Floating Point” on page 29 for details on the Oak language’s floating
point implementation.

3.3 Boolean Types

The boolean type is used for variables that can be either true or false, and for
methods that return true and false values. It’s also the type that is returned by
relational operators such as >.

Boolean values are not numbers and can’t be converted into numbers by casting.

Confidential Oak Language Specification 9

 3 Types © firstperson, inc.

2.4.3 Boolean Literals

The boolean type has two literal values: true and false. See “Boolean Types” on
page 9 for more information on boolean values.

2.4.4 Character Literals

Character literals are currently
implemented much like in C. A character literal is a character (or group of characters representing a single
When Unicode support is im- character) enclosed in single quotes. Characters have type char and are drawn
plemented, escape sequences from the Unicode character set (see “Character Types” on page 10).
will change.

2.4.5 String Literals

A string literal is zero or more characters enclosed in double quotes. Each string
literal is implemented as a String object (not as an array of characters).

2.5 Operators and Miscellaneous Separators

The following characters are used in Oak source code as operators or separators:

+ – ! % ^ & * | ~ / > < ( ) { } [ ] ; ? : , . =

In addition, the following character combinations are used as operators:

++ –– == <= >= != << >> >>> += –= *= /= &= |= ^=

%= <<= >>= >>>= || &&

For documentation of each operator, see “Operators” on page 22.

3 Types
Every variable and every expression has a type. Type determines the allowable
range of values a variable can hold, allowable operations on those values, and the
meanings of the operations. A number of built-in types are provided by the Oak
language. Programmers can compose new types using the class and interface
mechanisms (see “Classes” on page 10 and “Interfaces” on page 18).

The Oak language has two kinds of types: simple and composite. Simple types are
those that cannot be broken down; they are atomic. The integer, floating point,
boolean, and character types are all simple types. Composite types are built on
simple types. The Oak language has three kinds of composite types—arrays,
classes, and interfaces. Simple types and arrays are discussed in this section.

3.1 Integer Types

Currently, all integer types Integers in the Oak language are similar to those in C and C++, with two
act like 32-bit, 2’s comple- exceptions: all integer types are machine independent, and some of the traditional
ment, signed integers.
definitions have been changed to reflect changes in the world since C was

8 Oak Language Specification Confidential

 2 Lexical Issues © firstperson, inc.

2.3 Keywords

The ushort, Cstring, The following identifiers are reserved for use as keywords. They must not be used
string, and unsynchro- in any other way.
nized keywords are obso-
lete. boolean Cstring goto protected throw
The protect and unprotect break default if public transient
keywords are subject to byte do import return try
change. case double instanceof short unprotect
catch else int static unsynchronized
enum isn’t implemented
yet. char enum interface string ushort
class final long super void
instanceof might become a clone finally new switch volatile
method instead of a key- const float private synchronized while
word. continue for protect this

2.4 Literals

Literals are the basic representation of any integer, floating point, boolean,
character, or string value.

2.4.1 Integer Literals

Integers can be expressed in decimal (base 10), hexadecimal (base 16), or octal
(base 8) format. A decimal integer literal consists of a sequence of digits
(optionally suffixed as described below) without a leading zero (0). If an integer
literal begins with 0x, it is interpreted as a hexadecimal integer. If a nonzero literal
begins with 0, it is interpreted as an octal integer. Hexadecimal integers can
include digits (0-9) and the letters a-f and A-F. Octal integers can include only the
digits 0-7.
Type determination is not im-
plemented yet. Forcing literals The type of an integer literal is the narrowest integer type that it fits in (see
to be long is not implemented “Integer Types” on page 8). A literal can be forced to be long by appending an L
yet. or l to its value.

2.4.2 Floating Point Literals

A floating point literal can have the following parts: a decimal integer, a decimal
point (“.”), a fraction (another decimal number), an exponent, and a type suffix.
The exponent part is an e or E followed by an integer, which can be signed. A
floating point literal must have at least one digit, plus either a decimal point or e
(or E).

Double precision, NaN, As described in “Floating Point Types” on page 9, the Oak language has two
Inf, and the type suffixes floating point types: float (IEEE 754 single precision) and double (IEEE 754
are not implemented yet.
double precision). You specify the type of a floating point literal as follows:
2.0d or 2.0D double
2.0f or 2.0F or 2.0 float

Specifying too many significant digits for a single precision literal is an error.

Confidential Oak Language Specification 7

 2 Lexical Issues © firstperson, inc.

2 Lexical Issues
During compilation, the characters in Oak source code are reduced to a series of
tokens. The Oak compiler recognizes five kinds of tokens: identifiers, keywords,
literals, operators, and miscellaneous separators. Comments and white space such
as blanks, tabs, and line feeds are not tokens, but they often are used to separate
tokens.
Unicode source files aren’t
allowed yet because Oak programs are written using the Unicode character set, or some character set
there’s no editor/develop- that is converted to Unicode before being compiled.
ment environment to gen-
erate them. Instead, ASCII
input is accepted.
2.1 Comments

The Oak language has four kinds of comments:

// text All characters from // to the end of the line are

ignored.

/* text */ All characters from /* to */ are ignored.

See the oak(1) man page /** text */ Like /*...*/, except that these comments are treated
for information on auto- specially when they occur immediately before any
matically generating docu-
mentation. declaration or when they occur on the same line as a
declaration (even if after it). These comments
indicate that the enclosed text should be included in
automatically generated documentation as a
description of the declared item.

//* text Like //, except that these comments, like /**...*/,
indicate text to be included in automatically
generated documentation. Any subsequent //
comments with no intervening code are included in
the automatically generated documentation for the
declared item.

2.2 Identifiers

Unicode identifiers aren’t Identifiers must start with a letter, underscore (“_”), or dollar sign (“$”);
implemented yet. Instead, subsequent characters can also contain digits. For the part of Unicode that
identifiers are ASCII and
follow the C rules. overlaps ISO-Latin-1, letters are the characters “A” through “Z”, “a” through “z”,
and all the accented letters. Other characters valid after the first letter of an
identifier include every character except those in the segment of Unicode reserved
for special characters.

Thus “garçon” and “Mjølner” are legal identifiers, but strings containing
characters such as “¶” are not.

6 Oak Language Specification Confidential

 © firstperson, inc.

Oak Language Specification

This document is a preliminary specification of the Oak language. Both the
specification and the language are subject to change. When a feature that exists in
both Oak and ANSI C isn’t explained fully in this specification, the feature should
be assumed to work as it does in ANSI C.

1 Program Structure
The source code for an Oak program consists of one or more compilation units.
Each compilation unit can contain only the following (in addition to white space
and comments):
• a package statement (see “Packages” on page 19)
• import statements (see “Packages” on page 19)
• class declarations (see “Classes” on page 10)
• interface declarations (see “Interfaces” on page 18)

This restriction is not yet Although each Oak compilation unit can contain multiple classes or interfaces, at
enforced by the compiler, most one class or interface per compilation unit can be public (see “Classes” on
although it’s necessary for
efficient package importa- page 10).
tion (which is documented
in “Packages” on page 19). When Oak source code is compiled, the result is Oak bytecode. Oak bytecode
consists of machine-independent instructions that can be interpreted quickly by
the Oak runtime system.

Implementation Note: In the current Oak implementation, each compilation

unit is a file with an “.oak” suffix.

1.1 Starting Oak Programs

When an Oak program is executed, the interpreter must determine which method
to execute first. It does so by executing the class method main() in the class that’s
In the UNIX implementa- specified when the interpreter is invoked. The main() method must have the
tion, the classname is spec- following definition:
ified as follows:
oaki ClassName args
public static void main(String arguments[]) {
/* startup code goes here */
}

At least one class per program must implement the main() method.

Confidential Oak Language Specification 5

 © firstperson, inc.

6 Packages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
6.1 Specifying a Compilation Unit’s Package . . . . . . . . . . . . . . . . . . . . . 20
6.2 Using Classes and Interfaces from Other Packages. . . . . . . . . . . . . 20

7 Assertions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
7.1 Constraints on Instance Variables and Methods . . . . . . . . . . . . . . . 21
7.2 Preconditions and Postconditions . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

8 Expressions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
8.1 Operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
8.1.1 Operators on Integers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
8.1.2 Operators on Boolean Values . . . . . . . . . . . . . . . . . . . . . . . . 23
8.1.3 Operators on Floating Point Values. . . . . . . . . . . . . . . . . . . 23
8.1.4 Operators on Strings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
8.1.5 Operators on Objects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
8.2 Casts and Conversions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

9 Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
9.1 Declarations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
9.2 Expressions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
9.3 Control Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
9.4 Exceptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
9.4.1 The finally Statement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
9.4.2 Asynchronous Exceptions. . . . . . . . . . . . . . . . . . . . . . . . . . . 28

10 Garbage Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

A Appendix: Floating Point. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

A.1 Special Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
A.2 Binary Format Conversion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
A.3 Ordering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
A.4 Summary of IEEE-754 Differences . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

iv Contents Confidential
 © firstperson, inc.

Contents

1 Program Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.1 Starting Oak Programs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2 Lexical Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.1 Comments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.2 Identifiers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.3 Keywords . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.4 Literals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.4.1 Integer Literals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.4.2 Floating Point Literals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.4.3 Boolean Literals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.4.4 Character Literals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.4.5 String Literals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.5 Operators and Miscellaneous Separators . . . . . . . . . . . . . . . . . . . . . . 8

3 Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.1 Integer Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.2 Floating Point Types. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.3 Boolean Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.4 Character Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
3.5 Arrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

4 Classes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
4.1 Instance Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
4.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
4.3 Overriding and Overloading Methods . . . . . . . . . . . . . . . . . . . . . . . 13
4.4 Used before Set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
4.5 Class Variables and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
4.6 Constants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
4.7 Volatile Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
4.8 Transient Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4.9 Final Classes and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4.10 Access to Variables and Methods. . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4.11 Synchronized Methods and Blocks . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4.12 Constructors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
4.13 Order of Declarations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

5 Interfaces. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
5.1 Interfaces as Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
5.2 Constants in Interfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
5.3 Combining Interfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

Confidential Oak Language Specification iii

  1994 FirstPerson, Inc. All Rights Reserved.
100 Hamilton Avenue, Palo Alto California 94301 U.S.A.
This product and related documentation are protected by copyright and distributed under licenses restricting its use, copying,
distribution, and decompilation. No part of this product or related documentation may be reproduced in any form by any
means without prior written authorization of FirstPerson and its licensors, if any.
Third-party font software in this product is protected by copyright and licensed from Sun’s Font Suppliers.
RESTRICTED RIGHTS LEGEND: Use, duplication, or disclosure by the United States Government is subject to the restrictions
set forth in DFARS 252.227-7013 (c)(1)(ii) and FAR 52.227-19.
The product described in this manual may be protected by one or more U.S. patents, foreign patents, or pending applications.
TRADEMARKS
FirstPerson, the FirstPerson logo, the FirstPerson agent, Sun, Sun Microsystems, Sun Microsystems Computer Corporation,
the Sun logo, the SMCC logo, are trademarks or registered trademarks of Sun Microsystems, Inc. UNIX and OPEN LOOK are
registered trademarks of UNIX System Laboratories, Inc., a subsidiary of Novell, Inc.. All other product names mentioned
herein are the trademarks of their respective owners.
All SPARC trademarks, including the SCD Compliant Logo, are trademarks or registered trademarks of SPARC International,
Inc. SPARCstation, SPARCserver, SPARCengine, SPARCworks, and SPARCompiler are licensed exclusively to Sun
Microsystems, Inc. Products bearing SPARC trademarks are based upon an architecture developed by Sun Microsystems, Inc.
The OPEN LOOK® and Sun™ Graphical User Interfaces were developed by Sun Microsystems, Inc. for its users and licensees.
Sun acknowledges the pioneering efforts of Xerox in researching and developing the concept of visual or graphical user
interfaces for the computer industry. Sun holds a non-exclusive license from Xerox to the Xerox Graphical User Interface,
which license also covers Sun’s licensees who implement OPEN LOOK GUIs and otherwise comply with Sun’s written license
agreements.
X Window System is a trademark and product of the Massachusetts Institute of Technology.
THIS PUBLICATION IS PROVIDED “AS IS” WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESS OR IMPLIED,
INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY, FITNESS FOR A
PARTICULAR PURPOSE, OR NON-INFRINGEMENT.
THIS PUBLICATION COULD INCLUDE TECHNICAL INACCURACIES OR TYPOGRAPHICAL ERRORS. CHANGES ARE
PERIODICALLY ADDED TO THE INFORMATION HEREIN; THESE CHANGES WILL BE INCORPORATED IN NEW
EDITIONS OF THE PUBLICATION. FIRSTPERSON, INC. AND/OR SUN MICROSYSTEMS, INC. MAY MAKE
IMPROVEMENTS AND/OR CHANGES IN THE PRODUCT(S) AND/OR THE PROGRAM(S) DESCRIBED IN THIS
PUBLICATION AT ANY TIME.

Please
Recycle
Oak Language Specification

FirstPerson, Inc.
100 Hamilton Avenue
Palo Alto, CA 94301
U.S.A.