Hi,
I once read a paper about a fast matcher written in Lisp. I browsed
the AI repository and did not find it. I can't remember the name of
that system, I think it had a three-letter acronym. It must be quite
old.
The trick used in this paper was to use macros so that while writing a
function that would try to test if a given expression matches, the
expression that I write would be expanded to Lisp code that could then
be compiled. This is very useful when the patterns are known.
Any pointers to this article?
Jo"rg Ho"hle.
············@gmd.de ······@zeus.gmd.de
http://zeus.gmd.de/~hoehle/amiga.html
······@zeus.gmd.de (Joerg Hoehle) writes:
>I once read a paper about a fast matcher written in Lisp. I browsed
>the AI repository and did not find it. I can't remember the name of
>that system, I think it had a three-letter acronym. It must be quite
>old.
>The trick used in this paper was to use macros so that while writing a
>function that would try to test if a given expression matches, the
>expression that I write would be expanded to Lisp code that could then
>be compiled. This is very useful when the patterns are known.
>Any pointers to this article?
Probably not what you had in mind, but Norvig's book (Pardigms of AI
programming) has some material about compiled matchers.
Martin
--
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Martin Cracauer <········@wavehh.hanse.de> http://cracauer.cons.org
In article <······················@wavehh.hanse.de>,
········@wavehh.hanse.de (Martin Cracauer) wrote:
> ······@zeus.gmd.de (Joerg Hoehle) writes:
>
> >I once read a paper about a fast matcher written in Lisp. I browsed
> >the AI repository and did not find it. I can't remember the name of
> >that system, I think it had a three-letter acronym. It must be quite
> >old.
>
> >The trick used in this paper was to use macros so that while writing a
> >function that would try to test if a given expression matches, the
> >expression that I write would be expanded to Lisp code that could then
> >be compiled. This is very useful when the patterns are known.
>
> >Any pointers to this article?
>
> Probably not what you had in mind, but Norvig's book (Pardigms of AI
> programming) has some material about compiled matchers.
See
ftp://ftp.netcom.com/pub/hb/hbaker/Prag-Parse.html (also .ps.Z).
It shows you how to compile a complete recognizer/parser for Common
Lisp lambda functions, with all their &hairballs.