Re: URL +1, LSID -1

On 7/13/07, Xiaoshu Wang <wangxiao@musc.edu> wrote:

> Interestingly, when RDF was first developed some 8-9 year ago, it was
> intended for the framework of metadata.  "The solution proposed here is
> to use /metadata/ to describe the data contained on the Web." (From the
> introduction of the 1999 recommendations).  I wonder why W3C has
> abandoned such "phrasing", perhaps W3C intended to make RDF a more
> universal data model.

Maybe you haven't been paying attention to what W3C has been up to
since 1999 (this would be odd for someone subscribed to a w3.org
list), but this is exactly what happened, starting about the time the
XML recommendation came out: W3C started advocating HTTP URIs as
identifiers for everything, with the pun of using one as a locator
fully embraced (as opposed to violently rejected, as the LSID
community claimed to do... don't get me started on the question of
whether LSID and HTTP URI are any different with respect to their
locator-ness and identifier-ness...). Semantic web amplifies this
practice enormously - you may have noticed that e.g. predicates such
as rdf:type, without which it is almost impossible to use the semantic
web, are identified by http URIs, so you cannot reject this practice
without rejecting all of the semantic web recommendations.

I am sorry I have been using the term "metadata". What I really mean
is "description" or "defining description" represented in some RDF
syntax. The world uses "metadata" to mean description of data
(http://www.google.com/search?q=define:metadata), but I usually mean
description of anything. I think I may have been brainwashed by the
LSID spec, which uses "metadata" to mean any kind of description of
anything identifiable by an LSID. (Recall that an LSID may identify an
"abstract concept", not just a piece of data, and any LSID lacking a
version component has no data, only so-called "metadata".)

Jonathan

Received on Friday, 13 July 2007 14:37:43 UTC