Hi there,
I'd like to offer my own perspective on the GILS/DC question. My company is
and has been involved in several projects working with GILS in different
applications with no relationship to the US Government - ranging from
distributed catalogs of WWW resources over environmental protection efforts
to legislative information resources. The information model used in GILS is
a superset of the base DC elements, and the GILS profile has been tracking
the development of the DC elements, practically since the first meeting in
Dublin, Ohio to my knowledge. Indeed, most of the projects we work in lean
heavily on the DC when they aim to create common data models to be used by
different organisations - each with their own "legacy" data models.
The DC does its job very well here - as a sort of common semantic reference
model that most people can ake some sense of. GILS is more than just a data
model. It specifies exactly how one might search a database using the
Z39.50 protocol, making those data available to stand-alone clients,
parallel-searching WWW gateways, or traditional library systems. The
contents of the databases might be generated out of an existing database,
mapping their contents into the DC/GILS elements, or they might be created
by a WWW harvesting robot crawling through web-pages with embedded DC
metadata (as Sigfrid Lundberg suggests).
Again: the value of GILS to us is that it allows you to INDEX and SEARCH
databases of "DC-like" records in a well-defined way, and using software
which is commonly available. Because GILS is an open standard, we can throw
together software and resources without worrying too much about the
underlying software.
Looking at GILS and DC exclusively as two different data models: The GILS
one is significantly richer than the base DC set-of-fifteen, but it is
*arguably* less extensible than "structuralist" DC (is that the right
word?). We have found that the GILS set of elements is generally more
useful than plain DC - particularly because it is also good at describing
non-electronic resources - books, institutions, or individual persons.
But in sum - the DC and GILS don't compare directly. GILS is an excellent,
standard mechanism for searching remote DC-based databases. GILS does have
a data model which is more complex than the simple list of DC elements, but
while we wait for a profile specifically for using the DC in Z39.50, it is
a powerful tool. I can't answer for the status of GILS in the US
government. We've always used GILS strictly because it meets a technical
requirement.
--Sebastian
--
Sebastian Hammer <[log in to unmask]> Index Data
Ph.: +45 3536 3672 <http://www.indexdata.dk> Fax: +45 3536 0449
|