metadata in HTML pages

[Lee Jaffe]

Dear List:

I'm hoping both for some specific answers which will help me resolve some
questions I have and also perhaps stimulate discussion which may be useful
generally.  I've been working to incorporate metadata into our library's
Web pages.  Some of the information required -- such as a <title> tag --
falls easily into the "no-brainer" category,  Some tags help search engines
find, categorize and display your pages better. There are some other mechanical
or functional metatags with local benefits.  For instance, we use ht://DIG
as our local search engine and there are metatags which allow us to take
advantage of added indexing features of that system.

To help local Web authors incorporate recommended metadata in their pages,
we've been developing online forms which can kick out the required coding.
In the midst of discussing and demonstrating these features, someone asked
me whether we shouldn't be using Dublin Core, instead or as well.  I have to
admit, I was stumped.  I know DC and I've used in within some of my own
projects but it doesn't jump out at me as a solution for this sort of
application (i.e., everyday library Web pages).  In other words, what would
be the advantages of coding DC (or even another metadata scheme) into our
pages?  Are there common search engines which can take advantage of any
of these metadata schemes?  Or other tools which do something interesting
with this information.

Within the academic, research and library worlds, I see that DC and other
metadata schemes are taken very seriously.  I and expect that within certain
projects -- XML and SGML encoded documents and digital archives organized
within a special retrieval and browsing interface -- these can be critical. 
But what about the average Web page, such as a list of departments, phone
numbers, and names?  Is there any point in using one of these higher-end
metadata schemes generally? 

I've asked this question in another form some time ago.  What I actually
asked then was why wasn't the library community putting more effort into
getting a metadata standard implemented by the search engines.  I didn't
really get a satisfactory answer and I haven't seen any developments tending
towards making all our metadata efforts worthwhile generally. 

In fact, part of what prompted me to raise this question is the introduction of
Metadata Object Description Schema (MODS) http://www.loc.gov/standards/mods/
yet another library-based metadata standard.  MODS appears to challenge
existing standards, noting that it is "richer than Dublin Core" and "more end
user oriented than the full OAI MARC XML schema."  OK, so what now?  What if
I'd already started implementing one of the existing schemas in my library's
Web site, assuming that I didn't believe it to be a useless exercise in the
first place?  If LOC is right, I should be replacing DC with MODS coding.
Are the reasons that compelling?

I return to my earlier question:  Wouldn't it be more useful if the library
community spent its time and energy getting the search engine community to
adopt a  metadata schema as a retrieval option than spending our time spinning
out more and more refined standards that won't be used? 

I may be reading this whole business wrong.  However, I'm far from new to the
metadata question and while I understand the application within certain situ-
ations, I'm willing to accept the possibility that I've missed something impor-
tant in how metadata fits into the wider context of the Web.

Brickbats and kudos equally welcomed,

-- Lee Jaffe
-- 
===============================================================
Lee David Jaffe                  Library Computing & Network Services 
168B McHenry Library                                 ldjaffe at cats.ucsc.edu
University of California                831.459.3297  fax: 831.459.8206
Santa Cruz, California 95064           http://people.ucsc.edu/~ldjaffe