graph reduction

In computer science, graph reduction implements an efficient version of non-strict evaluation, an evaluation strategy where the arguments to a function are not immediately evaluated. This form of non-strict evaluation is also known as lazy evaluation and used in functional programming languages. The technique was first developed by David Turner in 1976.
See more at Wikipedia.org...

A technique invented by Chris Wadsworth where an expression is represented as a directed graph (usually drawn as an inverted tree). Each node represents a function call and its subtrees represent the arguments to that function. Subtrees are replaced by the expansion or value of the expression they represent. This is repeated until the tree has been reduced to a value with no more function calls (a normal form).
In contrast to string reduction, graph reduction has the advantage that common subexpressions are represented as pointers to a single instance of the expression which is only reduced once. It is the most commonly used technique for implementing lazy evaluation.