At Thu, 31 May 2007 10:42:57 -0700, Greg Meredith wrote:
BTW, i think this could have a lot of bang-for-buck because the literature i read exhibited two basic features:
- the "standard" treatments (even by CS-types) are decidedly not compositional - the people in the field who face industrial strength csp problems report that they have to take compositional approaches because the problems are just too large otherwise (both from a human engineering problem as well as a computational complexity problem)
This paper describes a non-monadic, compositional method for solving CSPs: http://www.cse.ogi.edu/PacSoft/publications/2001/modular_lazy_search_jfp.pdf There is also the LogicT monad transformer: http://okmij.org/ftp/Computation/monads.html j.