Value types
Syntax
Built-in types often come equipped with their own syntax for applying operations. For example, numbers can have infix binary operators such as +, –, etc., and prefix unary operators such as – and the square root symbol; Booleans have infix binary operators representing and, or, etc., and the prefix unary operator not. Each of these sets of operators has its own natural precedence for bracketing.
Our scheme for operator syntax allows operator signatures to be defined as prefix, infix or postfix using a positional notation. For example:
- the prefix unary minus for Integer has the signature
Integer : Integer;
- the prefix not for Boolean has the signature
not Boolean : Boolean;
- the infix less-than has the signature
Integer<Integer : Boolean;
- the postfix squared uses postfix dot notation as in
Integer.squared : Integer.
Only operators defined using the postfix dot notation may take additional parameters, as in Rectangle.contains(Point): Boolean. This notation is the same as that used for object navigation expressions and message-sending.
The following operator precedence rules apply, from highest to lowest, to all expressions (including those involving objects):
- non-alphabetic unary prefix;
- non-alphabetic unary postfix;
- alphabetic unary prefix;
- alphabetic unary postfix (dot notation with no parameters);
- infix binary multiplicative;
- infix binary additive;
- other infix binary;
- dot notation with parameters;
- other non-alphabetic infix;
Parentheses ( ) may be used to override these rules, in the normal way. When operators of the same precedence appear unbracketed together, the left-most takes precedence.
Boolean
Literals
{true, false}
Type specification
Character
Literals
{ @a, @b, @c, ..., @A, @B, @C, ...} : characters are prefixed by @ signs.
Type specification
The only operation defined on characters is equality: @a=@a, etc.
Symbol
Literals
{ %symbol, ... } : symbols are prefixed by percent signs.
Type specification
The only operation defined on symbols is equality: %abc=%abc, etc.
Number
Literals
{1.0, 1.1, 123456789.987654321, ...} : arbitrary-precision rational numbers specified using decimal notation.
Note that {0, 1, 2, ...} are Integer literals, see below: Integer is a sub-type of Number, so these literals also denote Number values.
Note that Number values can also be denoted by dividing two Integer values, for example 22/7.
Type specification
Integer
Literals
{...,-2,-1,0,1,2,3...} : the normal integer numerals.
Type specification
Integer is a sub-type of Number. This means that an Integer value can be used wherever a Number value is expected. This has the following implications:
- The set of Integer values is a subset of the Number values.
- The Number operations are inherited or overridden by Integer.
- Any operation overridden by Integer agrees, in the sense that the result obtained by applying it to Integer values is the same result as obtained by applying the overridden version to Number values.
Integer sub-ranges
The syntax of sub-ranges of Integer is lower..upper, lower being the lower bound and upper the upper (e.g. 1..10).
String
Literals
{'hello world', ... } : strings are contained in single quotes. The empty string is denoted by ''.
Type specification
Comments, corrections to
jdaniels@cix.co.uk