From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!uwm.edu!csd4.csd.uwm.edu!litow Tue Jan 21 09:27:21 EST 1992
Article 2909 of comp.ai.philosophy:
Newsgroups: comp.ai.philosophy
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!uwm.edu!csd4.csd.uwm.edu!litow
>From: litow@csd4.csd.uwm.edu (Bruce E Litow)
Subject: table look-up
Message-ID: <1992Jan20.142740.13327@uwm.edu>
Followup-To: Intelligence=retrieval,...  
Summary: Tables need not look like the desired data ...
Originator: litow@csd4.csd.uwm.edu
Keywords: tables, uniformity, content/structure of tables
Sender: news@uwm.edu (USENET News System)
Organization: Computing Services Division, University of Wisconsin - Milwaukee
Date: Mon, 20 Jan 1992 14:27:40 GMT
Lines: 27

Mark Hopkins 
(uwm.edu!csd4.csd.uwm.edu!markh Mon Jan 20 07:29:45 CST 1992)
points out that `good' tables for look-up are built by:

> By squeezing out all of its redundancy.

> Stupid programs are called stupid
> because they take so long to access the information they need (to make a move
> or solve a problem or whatever).  People too, I guess.

> The question you should be asking is: how does one STORE such a table in the
> first place.  Now storing information, that's what learning is.  

I agree with this and would like to remind readers that very small and very
efficient tables may have a structure that is far removed from the sought
after data. For a famous example, there is a n^O(1) look-up
table for all primes up to length n. Look-up time is basically O(n). 
But building this table is another matter (and recognizing it if it
fell from oracle land). More generally non-uniform circuit families
can be regarded as very efficient tables whose generation is indeed
another story.

-- 
Bruce Litow
Computing Services Division
P.O.Box 413, University of Wisconsin-Milwaukee, 53201
414 229 6431    litow@csd4.csd.uwm.edu


