Subject: Re: XML and GILS
Carl Hage (carl@chage.com)
Date: Sat, 28 Mar 1998 14:06:17 -0800
Message-Id: <199803282227.QAA29431@rgate.ricochet.net> From: "Carl Hage" <carl@chage.com> To: gils@cni.org Date: Sat, 28 Mar 1998 14:06:17 -0800 Subject: Re: XML and GILS
On Sat, 28 Mar 1998, Eric J. Miller <emiller@oclc.org> wrote:
>
> Carl Hage <carl@chage.com> wrote:
> >
> > RDF has had the same problem as XML-- lack of semantics for
> > attributes. If you ask me, 99% of the value of GILS is in the
> > definitions of the attributes and controlled vocabularies. ...
>
> Carl,
>
> The W3C RDF Working Groups [0] are in essence all about how to
> represent and define semantics. As such, I don't understand your
> statement.
Sorry, I probably shouldn't have made such a simplistic statement.
Also I wrote in based on recollections of an older release of the RDF
documents. (I read a newer one, but not the very latest.) Also, there
is secret work not accessable to me, so I can't comment on that.
Many of my earlier critisisms have been addressed in subsequent
versions of the RDF specifications, and I see placeholders for other
important pieces, e.g. online access to the metadata and schema
definitions.
RDF provides a means to exchange definitions of arbitrary objects,
with properties and values. As such, it is useful to exchange metadata
including GILS.
The problem is that if everyone uses XML/RDF to exchange different
sets of objects and attributes, then there is effectively no
communication between applications. I think more effort should be
directed towards standardizing the property/attribute definitions. The
DC is too primitive, and the flat card-catalog/book model of MARC/GILS
needs to be extended to better fit the new object-oriented database
models, and also to represent the complex hierarchic structures of
organizations and information on the net. I think the work Eliot did
on the BSR mapping, and the work others have done on mapping GILS to
SGML/XML/RDF are moving GILS in the right direction, and address the
need to standardize the attributes.
I think the most needed area for future work is standardzation of
definitions and access methods for controlled vocabularies. A postal
address to a copyrighted paper thesaurus isn't very helpful. The lack
of access to CVs and lack of tools to manipulate CVs and integrate
them is a severe problem.
Here's a case in point. I needed to convert some GIS data in
state-plane coordinate units, and needed the conversion constants. So
I go to the FGDC metadata standard, and it gives a FIPS publication
and mailing address of the NIST. I go to www.nist.gov, but they don't
have standards in thier menu! I use thier search engine on "FIPS" and
get 0 hits. I used AltaVista to search the planet, and found the FIPS
in an obscure place inside NIST (div897 of the info tech lab), but
the referenced FIPS was withdrawn. (I accidentially found another
standard that has some info that I needed for another project
though.) Because an electronically accessable definition of code
values was not used, FGDC now incorporates and references undefined
code values.
A few days ago, Eliot mentioned GILS is a FIPS and said it was
"available from the NIST". Unfortunately, he didn't supply a URL. Now
I know if one goes to www.nist.gov, they get 0 hits on the word FIPS,
but just for fun I thought I would try to find the GILS FIPS via
GILS at GPO. It found a few records with the word FIPS in it, and
found a FIPS record. The GILS record only gives the postal address
and telephone for purchasing paper and mag tape copies, and does not
contain the locations for net accessable versions. The GILS text is
misleading because it improperly implies (mentions) that online
versions are not available.
I think GILS should be altered to make URLs much more prominent-- a
URL should be given greater precidence than postal addresses and
telephone numbers. US Federal GILS policy should mandate inclusion of
URLs if electronic documents or orgnanization/contact sites exist,
even if outside the source agency. Also, URLs for documents should
be required to be specific to the document, not simply an agency home
page, sometimes accompanied by a paragraph of instructions for
navigating through a GUI.
There is a problem where agencies do not disclose locations of free
electronically available versions of thier documents since they sell
them. Sure, we have a policy that mandates creation of GILS records,
but we have a practice of omitting electronic references in these
records. If agencies do not make electronic versions directly
accessable, then they should be required to solicit locations of
outside sources and include those in thier GILS records. We need to
make sure the GILS records really locate the information.
To summarize, what I think limits GILS most are issues other than
physical data record formatting.
> [0] http://www.w3.org/RDF
> [1] http://www.w3.org/RDF/Group/Syntax
This is a private password protected URL.
> [2] http://www.w3.org/XML
> [3] http://www.w3.org/RDF/Group/Schema
--------------------------------------------------------------------------
Carl Hage C. Hage Associates
<mailto:carl@chage.com> Voice/Fax: 1-408-244-8410 1180 Reed Ave #51
<http://www.chage.com/chage/> Sunnyvale, CA 94086
This archive was generated by hypermail 2a16 : Tue Mar 23 1999 - 03:55:42 EST