Thesaurus as web site "browse engine"

Steve Morris stemor at info.SIMS.Berkeley.EDU
Thu May 30 13:53:16 EDT 1996


Greetings,

   I was wondering if anyone's been playing with using a thesaurus structure
as a navigational tool for a web site.  I'm currently trying to put a
bibliographical database (8300 records) up on the web with both query and
browse access.   My intention is to use the subject thesaurus pertaining to
the database (Intelligent Transportation Systems)  to provide a framework
for links to reverse chronological lists of citations to which the terms
have been assigned.  As it stands now, links to term/citations pages
originate from: 1) an alphabetical term index, 2) from thesaurus terms
occurring as assigned terms within citations, 3) from term/citations pages
of terms that are related to in the thesaurus structure.
(related,narrower,broader, use for)

A simplified example of what I am trying to do can be found at:

   http://www.lib.berkeley.edu/~path/        (look under "browse by ITS
thesaurus term")

At that site, since last year, I've been keeping a monthly recent additions
list for the 120-125 new records added each month.  I wrote a Perl script
that reads in the raw database records and as well as a text file holding
the thesaurus structure and then generates all the pages and links.  In the
case of the recent additions list, thesaurus terms only occur as links if
the term is used in one or more of that month's records--and I'm toying with
eliminating links from within citations in cases where the term is only used
once.

My main concern is that there might be too much "link noise."   On the
individual term index pages I've tried to cluster all of the links to
related, broader, and narrower terms into a compact <h5></H5> construct near
the top of the page.   If there are a lot of associated terms, it gets a bit
busy.
     I hope to mount the full database soon, but would like to get some
feedback about different ways to handle using the thesaurus as an access
tool.  So far it seems like the "browse engine" is taking 5 times as much
time to develop as the search engine.
     I also would heartily welcome any criticisms I could get on how I've
set up the "Recent Additions List" at the URL listed above (like everything
else, it's a work in progress).

Thanks,

Steve

------------------------------------------------------------------------------
Steve Morris, Technical Services Librarian
Harmer E. Davis Transportation Library
Institute of Transportation Studies, U.C. Berkeley
e-mail: smorris at library.berkeley.edu
phone: (510) 642-3604    fax: (510) 642-9180



More information about the Web4lib mailing list