Re: META tag standards, search accuracy

Nick Arnett (narnett@Verity.COM)
Mon, 14 Oct 1996 09:27:12 -0700


At 8:13 AM 10/14/96, Benjamin Franz wrote:

[snip]

>The HTML-WG
>(and the Dublin code) both are trying/tried force in their approaches to
>standard setting. Rather than going with the flow and riding the part of
>the flow headed in the direction they want to go they tried to build a dam
>to divert the entire river - and found that a river is a damn hard (and
>sometimes outright dangerous) thing to to try to stop.

Certainly a standards group should have the acumen to recognize when the
market is setting a standard and then endorse it, instead of trying to give
CPR to deceased equines.

I think you're making way too big a deal out of this. Keywords just aren't
that useful. If they were, we wouldn't need syntax and grammar in order to
communicate effectively. Language is complex and subtle; it's a mistake to
think that reducing it to a handful of nouns will make it significantly
more searchable. Keywords have been available in commercial search systems
for a couple of decades, but they haven't helped much.

Ontologies (which you might think of as keywords with semantic
relationships) do make a big difference, which is why the Dublin work is
important.

Nick

P.S. Where do I put the keywords in my e-mail message, which will show up
on the Web when it's archived?

KEYWORDS: KEYWORDS, SEARCH, FIND, INFORMATION, STANDARDS, HTML, WEB, DUBLIN
(the technology, not the city -- how do I specify?), SYNTAX, GRAMMAR,
SEMANTICS, SEMANTIC NETWORKS, KNOWLEDGEBASES, NATURAL LANGUAGE, PRECISION,
RECALL, ONTOLOGY, LIBRARIAN, SOFTWARE, COMPUTERS, INTERNET, ARNETT, ROBOTS,
META, INFOSEEK, ALTAVISTA, LYCOS, YAHOO,
VERITY,FULCRUM,DATAWARE,PLS,E-MAIL,RETRIEVAL,DOCUMENT,PARTRIDGE,PEAR,TREE