Re: WOFF Metadata - why reinvent the wheel?

On Wednesday, August 11, 2010, 1:09:52 AM, Toby wrote:

TI> Why does WOFF invent its own metadata format rather than reusing one of
TI> the W3C's existing ones? RDF/XML or RDFa could easily express the kind of
TI> data that WOFF seems to need.

TI> The example in appendix A could be expressed in RDF/XML as:

(lengthy example with four namespaces, omitted. Anyone fancy typing it all in again to their reply, without looking?)

As with many technologies, WOFF is pulled in opposing directions. On the one hand, the folks who are actually generating and consuming WOFF want something as simple and easy as possible. Indeed, plain text was a strong contender here.

On the other hand, I'm sure that metadata enthusiasts can dream up a sublimely rich and complex formulation worthy of many long discussions at an academic conference.

The current spec, which is implemented and is in use by foundries, is a lightweight XML encoding, fairly close to plain text, easy to type, human readable, and readilly transformed into an HTML rendering (e.g. for an infobox) using XSLT.

Then again, one person's data is anothers metadata.

Also, is a font designer really a 'freind of a freind' (foaf)? There are I am sure many possible ways to express this in RDF.

Note that the metadata is not used by the browser for rendering. Its  used by the foundry for licensing information, and to aid discoverability of the font data. And the presence of this feature was a strong factor in getting the font foundries to feel comfortable with WOFF. 

If an RDF representation is desired, then GRRDL would seem to be a reasonable way to produce it. I would have no opposition to adding a suitable GRDDL transform (either to the spec or to informative, supplementary materials). We had already talked about doing that with XSLT, to produce an infobox presentation.

-- 
 Chris Lilley                    mailto:chris@w3.org
 Technical Director, Interaction Domain
 W3C Graphics Activity Lead
 Co-Chair, W3C Hypertext CG

Received on Wednesday, 11 August 2010 10:12:28 UTC