Newsgroups: comp.ai.doc-analysis.misc,comp.ai.nat-lang
Path: cantaloupe.srv.cs.cmu.edu!bb3.andrew.cmu.edu!newsfeed.pitt.edu!gatech!swrinde!howland.reston.ans.net!newsfeed.internetmci.com!uwm.edu!spool.mu.edu!munnari.OZ.AU!news.hawaii.edu!pollarda
From: pollarda@Hawaii.Edu (Art Pollard)
Subject: Re: Software Knows to Deconstruct in Plain English
X-Nntp-Posting-Host: uhunix5.its.hawaii.edu
Message-ID: <DnCCBz.9nq@news.hawaii.edu>
Sender: news@news.hawaii.edu
Organization: University of Hawaii
References: <4ftflg$st9@la1.digilink.net> <4gdgnn$d2h@la1.digilink.net> <DOWDING.96Feb20182843@Gansett.ai.sri.com> <312C8F2B.1CA2@thomtech.com>
Date: Sun, 25 Feb 1996 16:55:11 GMT
Lines: 25
Xref: glinda.oz.cs.cmu.edu comp.ai.doc-analysis.misc:150 comp.ai.nat-lang:4605

In article <312C8F2B.1CA2@thomtech.com>,
Steve Finch  <sfinch@thomtech.com> wrote:
>
>Now a more impressive demo would be a 200,000 document collection (or
>even a 20,000 document collection for a start) with "dumb" ranked
>retrieval and Boolean systems by its side for comparison so that
>everyone can plainly see the advantage that their revolutionary
>breakthrough gives over the old, inferior technology.  So, c'mon
>Intelligent, what are you afraid of?


Yes, I vote for a test with a standard test collection.  (This could be 
for example CACM, etc.  Plenty of standard test collections may be found 
at Cornell Universities FTP site.)  Or, they could put up the Bible.  It 
has around 65,000 verses, is reasonably well known (compared to an AIDS 
database where nobody knows what is really in it), and the language is 
complex enough to provide a real test.  It is publically available as 
well. (Heck, I'll even supply it to them if they want a copy.)

I'm for a real test with text that people are familier with.  Otherwise, 
as far as I can tell, it is hype.

-Art


