ndd

[ Synopsis | Description | See also | Back to the NDD package documentation ]

This datatype holds the representation of a set of unbounded integer vectors sharing the same dimension. Such a representation can be created by calling one of the constructor functions provided by the NDD package. It can be unallocated by calling the function ndd_free.

Internally, an NDD is simply a finite-state automaton recognizing numbers. At the time an NDD is created, it is thus necessary to specify the base with respect to which the numbers are encoded, as well as the direction of this encoding (most or least significant digit first). For most applications, it is best to choose the base equal to 2 and to encode numbers most significant digit first. Some functions use a default direction for encoding numbers; one should not make assumptions about the value of this direction (other than supposing that it is consistent from one function call to another).