Date: 24 May 2005

Symbolic computing with compression of data structures: General observations, and a case study

* Final gross prices may vary according to local VAT.

Get Access


Data structures used by large symbolic computing systems tend to be fixed in form, and unable to change to reflect special symmetries or properties of individual computations. We examine the consequences of the view that the first step in a symbolic computation can be the analysis of the problem to be solved, to determine what is the most compact practical data structure for that problem. General principles of such an analysis are presented, and are then applied to a particular problem in differentiation which has caused difficulties in storage to some traditional large systems.