sbp: I'm currently surveying the current RDF-in-HTML options for deploying a CustomRdfDialect that I'm working on. GRDDL, RDFa, and eRDF seem to be the most prominent options.
sbp: At the moment I'm leaning towards RDFa, despite the disadvantage noted in this RFE, because GRDDL as a framework is too expensive: XSLT chaining is slow, and the multiple mechanisms make it even slower. The reinvention of @rel="stylesheet" is bad too, architecturally.
sbp: I'm not entirely sure that RDFa will be accommodating of my use case at the moment, but from what I've read so far it looks more likely than dislikely.
sbp: And then there's hGRDDL of course. I feel like a kid in a candy store... only all the candy is evil!
sbp: In other words, E_TOO_MANY_APPROACHES. Which means too many approaches to have to choose from, too many approaches to have to support, and the worst thing: too many approaches that can subsume one another in different directions. What are we making here, Ouroboros? Cf. my "It's important only to go as far with the meta-language thing as you have to." comment in RDF Stylesheets.
sbp: Mark Birbeck has responded thoughtfully, and hopefully I've managed to repay the courtesy in my own reply to him. The summary is that I think RDFa should make h:head/@profile="some-rdfa-id" a SHOULD and absolve user-agents of the horrors of processing documents which don't have it, in the default conformance case
sbp: See also Best Practices Issue: RDF Format Discovery which I just sent to the Semantic Web Deployment WG, since I think that this issue is clearly within their remit.
sbp: Wow, DanC raised basically exactly the same issue back in June 2007. He doesn't seem to have specified the solution in the same compromise-laden terms as I have though; I think he's asking to mandate @profile in document conformance, whereas I'm going for absolution in user-agent conformance.
sbp: At the moment I'm leaning towards RDFa, despite the disadvantage noted in this RFE, because GRDDL as a framework is too expensive: XSLT chaining is slow, and the multiple mechanisms make it even slower. The reinvention of @rel="stylesheet" is bad too, architecturally.
sbp: I'm not entirely sure that RDFa will be accommodating of my use case at the moment, but from what I've read so far it looks more likely than dislikely.
sbp: And then there's hGRDDL of course. I feel like a kid in a candy store... only all the candy is evil!
sbp: In other words, E_TOO_MANY_APPROACHES. Which means too many approaches to have to choose from, too many approaches to have to support, and the worst thing: too many approaches that can subsume one another in different directions. What are we making here, Ouroboros? Cf. my "It's important only to go as far with the meta-language thing as you have to." comment in RDF Stylesheets.
sbp: Mark Birbeck has responded thoughtfully, and hopefully I've managed to repay the courtesy in my own reply to him. The summary is that I think RDFa should make h:head/@profile="some-rdfa-id" a SHOULD and absolve user-agents of the horrors of processing documents which don't have it, in the default conformance case
sbp: See also Best Practices Issue: RDF Format Discovery which I just sent to the Semantic Web Deployment WG, since I think that this issue is clearly within their remit.
sbp: Wow, DanC raised basically exactly the same issue back in June 2007. He doesn't seem to have specified the solution in the same compromise-laden terms as I have though; I think he's asking to mandate @profile in document conformance, whereas I'm going for absolution in user-agent conformance.
sbp: 'The answer is, generally, "how long is a piece of string?", but in fact there are lots of cases in which we need to construct more specific answers.'
sbp: This is a writeup of some of the issues on the level above RDFa's ISSUE-38 and so on.