---------- Forwarded message ----------
Date: Mon, 2 Oct 2000 15:43:47 -0400
From: csyed <ad6509@wayne.edu>
Subject: Re: [Ref-Links] economic effects of link-based search engines on
               e-journals (William Y. Arms)

> William Y. Arms <wya@CS.CORNELL.EDU> wrote:

> The potential advantage of additional metadata is to improve the precision
> of searching.

I have been maintaining for some time that Metadata without a controlled
vocabulary is not likely to be effective. Now, I'm not so sure.

The LCSH for instance, are constructed on the premise of "warrant". If a term is
in general use, it eventually ends up as a subject heading. If authors select
their own keywords, engines will get confused. But building controlled
vocabularies has always been painfully slow. SO: let authors choose their own
descriptors - the terms most often used will be those indexed. There must be
something wrong with this notion, but I'm not sure just what... yet. Perhaps
it's because subject headings have additional fields, like "Java -- Programming
Language"... etc., (as opposed to coffee), which make them more powerful....

cbs

---
Chris Brown-Syed <ad6509@wayne.edu> <http://valinor.purdy.wayne.edu>
Ph: +1 313 577-0503.  Fax: +1 313 577-7563. Pager: +1 519 987-8409
Editor, Library & Archival Security. LIS Program, 106 Kresge Library,
Wayne State University, Detroit, MI, USA, 482023939