W3C home > Mailing lists > Public > www-rdf-interest@w3.org > October 2002

RE: TRANSFORM SVG2RDF

From: Danny Ayers <danny666@virgilio.it>
Date: Fri, 11 Oct 2002 17:29:06 +0200
To: "Charles McCathieNevile" <charles@w3.org>
Cc: "Emmanuel Pietriga" <epietriga@yahoo.fr>, "Diego Garcia" <diego_garcia77@hotmail.com>, <www-rdf-interest@w3.org>
Message-ID: <EBEPLGMHCDOJJJPCFHEFOEBBIBAA.danny666@virgilio.it>

Thanks Chaals, interesting to hear your take on this.

>But I am not sure why the metadata is hard to get inline - presumably you
>have a representation of the triples, and so you can add metadata
>elements to
>anything that represents an rdf:predicate ?

The difficulty was mostly down to my inability to keep a clear head ;-)
I got in a bit of a tangle between the Batik SVG tree and the Jena API - my
graph items could happily serialise to both, but I didn't figure a good way
of pulling out a single RDF description and inserting it into the DOM tree.
I'm sure this is doable in a reasonably neat way, but I also ran into more
generic app problems - e.g. Batik uses Crimson, Jena uses Xerces, and both
are pretty sizeable libraries. The way I'm thinking of moving forward with
this in my app is to probably drop Batik and Jena altogether. I'm not sure I
was using any SVG-specific parts of Batik anyway, so a regular DOM API
should do (although the CSS microparser would likely have been useful). I'm
contemplating writing something like a "Jena Lite", with only the minimum
needed to parse/retain/write RDF (if any of the Jena guys are reading this -
any thoughts?). The big advantage of this approach is it should allow for
the marshalling/unmarshalling of SVG+RDF right from the start, using a
common interface where the RDF & SVG descripions are implicitly tied
together. (hmm - I might have another look at Jena+Batik inlining, the same
interface could be used). The big disadvantage is the work involved...

An alternative would perhaps be to have the RDF play host to the SVG as
CDATA literals, but that wouldn't look so nice in a regular browser. A
stylesheet that could transform between the two approaches would make a fun
project!

Cheers,
Danny.
Received on Friday, 11 October 2002 11:39:39 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:56 GMT