Newsgroups: comp.lang.lisp
Path: cantaloupe.srv.cs.cmu.edu!das-news.harvard.edu!news2.near.net!MathWorks.Com!yeshua.marcam.com!zip.eecs.umich.edu!newsxfer.itd.umich.edu!isclient.merit.edu!msuinfo!harbinger.cc.monash.edu.au!news.cs.su.oz.au!metro!usage!news
From: Timothy James Menzies <timm@cse.unsw.edu.au>
Subject: Indexing onto disc structures
Message-ID: <1994Sep12.055223.14440@usage.csd.unsw.OZ.AU>
Sender: news@usage.csd.unsw.OZ.AU
Nntp-Posting-Host: dexter.orchestra.cse.unsw.edu.au
Organization: School of CompSci & Eng, Uni Of NSW, Oz
Date: Mon, 12 Sep 1994 05:52:23 GMT
Lines: 23

i've been puzzling how to build an sql system in lisp. peter norvig's
book has some excellent tips on how to build a generic query system
(using a prolog-like syntax). but on one point i find myself doubting
the words of the prophet....

norvig advocates the use of trie trees to index the facts and relations
in a database. i like his architecture: the trie trees do some rough
fast guessing which pass a set of candidates to an unifier which makes
the final decision. like it a lot. but my mind can't help wondering if
there is some simpler approach in the case where the facts are propositional,
contain no nested lists, and i have domain knowledge about the legal
values for each field (e.g. i will know that field 2 is day which is
one of mon, tues, wed, etc).

what i seek is a system that stores both its facts and indexes on disc.
suggestions anyone?

--
           Tim Menzies =timm@cse.unsw.edu.au| I'll play with it first
 ,-_|\     AI Lab, Computer Science,        | and tell you what it is
/     \    Univeristy of NSW, P.O. Box 1,   | later.
\_,-._* <- Kensington, Australia, 2033      |  -- Miles Davies
     v     ph:  +61-49-676-096              | 
