Lists Home |
Date Index |
Eric Hanson <email@example.com> writes:
> Michael Champion (firstname.lastname@example.org) wrote:
> > On Thu, 28 Oct 2004 21:02:12 +0000, Eric Hanson <email@example.com>
> > wrote:
> > > 1. There is no way to look up, discover and retrieve the
> library of
> > > resources that support with a namespace-qualified element. If you
> > > come across a piece of data, there may be hundreds of supporting
> > > resources like XSL transformations, schemas, xforms, text
> > > documentation, etc. We need a way to link the resources to the
> > > data.
> > It would seem that there at least a dozen ways that could be done
> > leveraging XML technolgies. The various flavors of RDDL
> (one based on
> > RDF, another on XLink) come to mind
> RDDL comes close, but it's not a distributed database, rather
> a single authoritative piece of documentation for a
> namespace. We need something where third parties can author
> resources and publish them to this distributed database, so
> they can be found by others.
This is discussion we just had on the "After XQuery, are we done?"
thread: starting from the question of how to glue separate documents
together. Given the twists and turns that thread took I wouldn't hold
out a lot of hope for solving this any time soon.
> > not to mention plain ol' XPath/XQuery
> What would be queried here?
> > and the various ways proposed in the web services world
> > (e.g. UDDI, WS-MetadataExchange).
> UDDI is in a similar spirit, but isn't it for locating web
> services? We're talking about supporting resources.
> > Arguably this is more or less
> > exactly what the Semantic Web would enable out of the box.
> Arguably this would enable the Semantic Web to get out of the
> box. :-) I mean, how does the Semantic Web enable this out
> of the box?
If all the components for the Semantic Web existed you'd arguably
already have universal relationship discovery. If you build all the
components needed for universal relationship discovery you arguably have
the Semantic Web. Two sides of the same coin; no need to debate which
one comes first since they'll both be boot strapping each other into
existing for a long time yet.
<snip>some implementation discussion and other related stuff</snip>
> So what's the point of having a universal language for data
> and a set of universal supporting resource types if there's
> no universal way to link the two together?
The first two are easy compared to the third?
> I took a shot at making something that would work a while
> ago: http://typekit.org/spec/ -- It's not much more than a
> strawman at this point -- not using RDF, just one database
> instead of distributed, etc.
> But it seems to me this is something the w3c should have been
> working on back in 2000. I think it needs to be as universal
> as our DNS system or XML itself, and look something like the
> XML-packaging WG charter.
The idea of w3c level ontology roots seems to pop up every once and a
while. I believe you're asking for something a little more local and a
little less heavy weight than ontology traversal to locate resource
descriptors so that you can query for document contents, but that's the
crux of the problem: anything universally applicable needs to solve
universal scale problems. At the moment I'd have to conclude that you
need to be careful what you ask for, you might get it...