Re: Meta Tag [and Knowledge Representation]

Davide Musella (davidmsl@ipo.tesi.dsi.unimi.it)
Mon, 20 Nov 1995 18:40:51 +0100 (MET)


From: davidmsl@ipo.tesi.dsi.unimi.it (Davide Musella)
Message-Id: <9511201740.AA68572@ipo.tesi.dsi.unimi.it>
Subject: Re: Meta Tag [and Knowledge Representation]
To: FisherM@is3.indy.tce.com (Fisher Mark)
Date: Mon, 20 Nov 1995 18:40:51 +0100 (MET)
Cc: www-html@w3.org (Mailing list di html)
In-Reply-To: <30AD0AFC@MSMAIL.INDY.TCE.COM> from "Fisher Mark" at Nov 17, 95 02:24:00 pm

> 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. 
[...]
Maybe this is the best way to catalogue a doc, but we are in a different
context.
If I want to do a full-text indexing, I must have the complete html file, if I
use the Meta tag, i can ask only for the header and the HEAD is normaly the 5%
of the complete file.
We have also to look at the net resources, and to the band width. 
The use of the Meta-Tag is not the best way, but only a good compromise among
CPU load, Net Load and a Good indexing system
Davide

-------
Davide Musella
davidmsl@anti.tesi.dsi.unimi.it