Re: need information on use of indexes

[Wayne Douglass <wayned -at- VERITY -dot- COM>]

At 03:04 PM 6/9/98 -0700, Janice Gelb wrote:

>>From a session I went to by Jared Pool:
>
>  * Indexes are still necessary online, since full text searches depend
>  on what the author included in the document (he said a search of
>  Smithsonian Magazine archives for "dinosaur" turned up an article
>  on the American steel industry but not one on tyrannasaurus rex).
>  Also, for web delivery, what about alternate spellings (like the
>  British "authorisation")?
>
I have more than one bone to pick with Jared Spool, but this statement can
be easily disproved by using the Smithsonian search page at
http://www.smithsonianmag.com:8765/. A search on "dinosaur" yields 22
matches with the steel industry story at #6. A search on "tyrannosaurus"
yields 4 matches, all but one included in the original list of 22. The
unique story is about a baby hippo named Chitwan, which the author compares
to Tyrannosaurus. The word dinosaur does not appear in the story, but the
word "dinosaurian" does, which suggests that the Smithsonian search engine
is relatively crude and does not do "stemming" to find variations of the
search term.

--Wayne
---------------------------------------------
Wayne Douglass        phone: 408-542-2139
Verity, Inc.            FAX: 408-542-2040
894 Ross Road        mailto:wayned -at- verity -dot- com
Sunnyvale, CA 94089    http://www.verity.com
---------------------------------------------