Preferred Listings

Michael Haseltine haseltin at ag.Arizona.EDU
Wed Jun 19 23:34:45 EDT 1996


The thinking this discussion is provoking is beginning to merge with some
other thinking I've been doing about search engines. It started with my
thinking about the fact that all the documents about writing good html talk
about the proper use of codes as containers, thinking about document
structure and the content of an element, rather than appearance. If this has
a purpose, it must be that there is a use for this structured way of
creating documents, but that as far as I can see, it's not being used.

The way I see it, if it were used as intended, search engines would give
results like the title and all heading 1 and 2 elements for a document
instead of title and the first n number of words on the page, most of which
is useless, in my experience. If they returned stuff in headings, then we'd
see the use of word spamming go away because it wouldn't be effective, and
we'd get search results that would be more meaningful.

It would also mean that searches could be performed like this: I want to see
all web pages that have bibliographies with citations for things that have
"desertification" in their titles.

This is the kind of searching I want to do. The broad, catch everything even
if it's misspelled, kind of search, with relevancy ranking, that nets 40,000
items does me very little good. Usually, I am looking for a very narrow set
of items.

If this is being provided by some search engines, I'm unaware of it and
would be happy to hear about it. What do other people think about this?
------------------------------------------------------------------------
Michael Haseltine -- Arid Lands Information Center, University of Arizona
haseltin at ag.arizona.edu



More information about the Web4lib mailing list