RE: Can RDF thrive in an XML-centric world?

>>It would be interesting to hear from the WG why the current RDF/XML hasn't
>>already "gone".
>
>My own view is that the WG has been constrained by the charter, and within
>that constraint has done an excellent job of cleaning up RDF/XML.

Yep, fair enough, it's easy to forget the organizational constraints from an
end user point of view.

>At least some of the problems with the M&S articulation of the RDF/XML
>syntax have been solved; for example the articulation in "RDF/XML Syntax
>(Revised)" of a grammar over XML infoset is a good thing, which fixes a
>number of bugs with M&S. As far as I recall, all the changes made by the WG
>to the grammar have been improvements.

(the last sentence makes me chuckle)
Most definitely - the fixes are great, and as a whole the documentation
suite makes RDF significantly more welcoming and easier to use than even
just a year ago.

>RDF/XML is also a deployed technology, so whatever else happens, we are
>likely to need continued support of it for a good while - cleaning up the
>specs helps this.

Agreed.

>But no amount of hard work on the part of the WG or the editor can make
>RDF/XML fit easily and comfortably into the XML world; the problems run too
>deep.

Also agreed. Graham's graph vs. tree point in the following mail can't be
avoided, but there is quite a bit of syntactical vinegar that could be
avoided.

>I agree that
>"a separate parallel (WG?) thread [sh]ould be spawned to work out a new
>syntax that had less surprises for XMLers"
>
>Hopefully the current WG is arriving at last call, which should allow such
>new work to be considered.

Graham's right about there already being other human-readable
representations (graph & n3), but these are pretty orthogonal from current
syntaxes in the industry (XML marketplace/whatever). The graph
representation depends on task-specific tools and the full n3 representation
(for all the benefits ascribed to it) if anything is likely to scare away
XMLers more than RDF/XML - no cosy angled brackets, more like the aftermath
of an ant battle. Ntriples is less confusing, but then a triple-only view
loses the big (graph) picture. Anyhow, I would hope & expect that this area
will at least be discussed, to see if it would be a productive use of
energy.

Cheers,
Danny.

Received on Monday, 4 November 2002 12:10:52 UTC