This file works together with pl/arrays.yap and arrays.h.

YAP supports a very simple notion of arrays Arrays may be allocated dynamically or statically:

o anonymous arrays are created during execution and allocated in the heap They have the lifetime of any other other heap object Any term can be an argument to a dynamic array

Dynamic arrays are named as a free variable and are initialized with free variables

o named arrays are created during execution but allocated in the code space They have the lifetime of an heap object Any term can be an argument to a dynamic array

Named arrays are named with atoms and are initialized with free variables

static arrays are allocated in the heap Their space is never recovered unless explicitly said so by the program Arguments to these arrays must have fixed size, and can only be atomic (at least for now)

Static arrays can be named through an atom They are initialized with []

Users create arrays by a declaration X array Arity If X is an atom A, then this it is a static array and A's the array name, otherwise X refers to a dynamic array

As in C, arrays start counting from 0

Users access arrays by a token X[I] or a[I], this token can appear anywhere within the computation, so a[2] = X[3*4] means that the second element of global array a should unify with the 12th element of array X The mechanism used to implement this is the same mechanism used to implement suspension variables

==== Representation:

Dynamic Arrays are represented as a compound term of arity N, where N is the size of the array Even so, I will not include array bound checking for now

|--------------------------------------------------------------|
| $ARRAY/N|....
|______________________________________________________________

Unbound Var is used as a place to point to

Static Arrays are represented as a special property for an atom, with field size and

A term of the form X[I] is represented as a Reference pointing to the compound term:

[]([I],X)