Re: Meta Tag [and Knowledge Representation]

Fisher Mark (FisherM@is3.indy.tce.com)
Fri, 17 Nov 95 14:24:00 PST


From: Fisher Mark <FisherM@is3.indy.tce.com>
To: www-html <www-html@www10.w3.org>
Subject: Re: Meta Tag [and Knowledge Representation]
Date: Fri, 17 Nov 95 14:24:00 PST
Message-Id: <30AD0AFC@MSMAIL.INDY.TCE.COM>


To add to Dan Connolly's comments...

I have recently learned enough about library science to make me glad I went 
into electrical engineering :).  This is a pretty sticky (and thereby 
interesting) problem!  You can certainly hack something together, but some 
study of the literature (as Dan noted) will likely be greatly useful.

If minimal index size is not a consideration, you should also spend time 
looking at full-text indexing systems rather than using keywords in META 
tags.  Back in 1971, Gerard Salton proved that a properly constructed 
full-text search system could provide better retrieval capabilities than the 
best keyword indexed system (ref. 
<URL:http://cs-tr.cs.cornell.edu:80/Dienst/UI/2.0/Describe/ncstrl.cornell%2f  
TR71-115?abstract=salton>, "A New Comparison Between Conventional Indexing 
(MEDLARS) and Automatic Text Processing (SMART)", Gerard Salton, Cornell 
Computer Science Technical Report TR71-115).  Unfortunately, the whole world 
has not yet caught up to the state of the art in 1971...

What looks to be a good book on these subjects (unfortunately still in my 
to-be-read pile) is Gerard Salton's _Automatic Text Processing_, 
Addison-Wesley, 1989, ISBN 0-201-12227-8.
======================================================================
Mark Leighton Fisher                   Thomson Consumer Electronics
fisherm@indy.tce.com                   Indianapolis, IN