[WEB4LIB] metadata in HTML pages

Chris Gray cpgray at library.uwaterloo.ca
Thu Jul 11 10:54:30 EDT 2002


We are somewhere on a continuum between two points, one is Web metadata as
only a gleam in someone's eye and Web metadata as universally deployed and
fully exploited.  The Web is gradually moving along this continuum and for
everyone it is a judgment call as to when you start implementing.  Or
perhaps a better analogy is this:  When are you going to buy a personal
computer and are you going to buy a Mac or PC or are you going to jump in
early and buy an Amiga?  When are you going to buy a VCR and are you going
to buy VHS or Beta?

"Standards are great! There are so many of them to choose from!"  You can
choose to enjoy the intellectual foment (and the ensuing chaos) or you can
despair at the time and effort expended on something that may eventually
have to be thrown away or redone, or do a bit of both.  After all,
libraries never have to re-catalog entire sections of their holdings,
right? ;)

I think the view expressed by Norm Medeiros in "Making Room for MARC in a
Dublin Core World" still holds true three years after he wrote it
<http://www.infotoday.com/online/OL1999/medeiros11.html>:  "The motivation
for AltaVista or Excite to adopt the Dublin Core syntax at present remains
questionable. A finalized W3C metadata RDF should spur search engine
companies into adopting the standard, and thus result in exact Dublin Core
element targeting. How this will play out remains to be seen.
Nevertheless, incorporating Dublin Core into library Web pages at present
can only help retrieval. Additionally, individual library search engines
can be crafted to target Dublin Core, thereby increasing retrieval for
users of that specific site."

Getting less philosophical, there is a list of DC tools and software at
<http://dublincore.org/tools/>, which includes IllumiNet Corpus that can
index DC in HTML.  I can't find any evidence that major search sites
properly parse DC, or any evidence that they look at <link rel="meta">
tags for external metadata or <head profile="URI"> attributes for metadata
profiles.

Perhaps the best advice is to pick the metadata standard that best suits
your project and that you find best.  We need less theorizing and standard
setting, and more practical experience on which to base theories and
standards.  The more people use it the better populated with metadata the
Web will become, which will make it worthwhile to build tools that
interpret and translate metadata.  I seem to be in quote mode today so
I'll end with, "If you build it, they will come."

Chris Gray
Systems Analyst
University of Waterloo Library

"general principle of robustness: be conservative in what you do, be
liberal in what you accept from others."  --Jon Postel, 1980, RFC 761, TCP





More information about the Web4lib mailing list