tag:blogger.com,1999:blog-8968818456594497310.post5586759348228165301..comments2009-03-02T06:18:22.676-08:00Joel Fernandeshttp://www.blogger.com/profile/04997674848378795829noreply@blogger.comBlogger3125tag:blogger.com,1999:blog-8968818456594497310.post-76122313057702518622009-03-02T06:18:22.676-08:002009-03-02T06:18:22.676-08:00The trick to getting fast pattern matching is to compile the patterns into the right VM instructions. There are papers describing how to do this. A good description is given in the book "The Implementation of Functional Programming Languages" by Simon Peyton-Jones. It is a very good book to get a feeling for implementing functional languages, unfortunately it is old and can be difficult to get hold of. The Erlang compiler uses the method described in the book.Robert Virdinghttp://www.blogger.com/profile/16187415127098159771noreply@blogger.comtag:blogger.com,1999:blog-8968818456594497310.post-57506704195022254762009-03-02T04:57:50.191-08:002009-03-02T04:57:50.191-08:00cl-unification is also nice. It supports full unification, and has a clean syntax. check it outJoel Agnel Fernandeshttp://www.blogger.com/profile/04997674848378795829noreply@blogger.comtag:blogger.com,1999:blog-8968818456594497310.post-11919555783063055982009-03-01T23:28:40.056-08:002009-03-01T23:28:40.056-08:00bpm appears to be a little bit unstable in development aspect, considering there is bpm2 and both have so-called changeable APIs, etc.<BR/><BR/>There is also another pattern matching library, called cl-match, which IMHO appears to be more lispish.<BR/><BR/>OTOH, I always wanted to learn how other languages (e.g. Erlang, OcaML, Haskell) packed with builtin pattern matching are able to perform pattern matching blazingly fast. One should investigate their DISASSAMBLE outputs.knthttp://www.blogger.com/profile/10946631855054644282noreply@blogger.com