SWBPD RDF-in-XHTML TF

 

Ralph: Pat Hayes' endorsement of CURIE was nice to see

Jeremy: yes, QName dependence on URI rather than IRI is going to be an issue

-> Re: [ALL] RDF/A Primer for review [Pat Hayes]

<jeremy> (On URI/IRI issue: Namespaces in XML (W3C Recommendation 1999)

<jeremy> An XML namespace is a collection of names, identified by a URI reference [RFC2396],

<jeremy> (not an IRI reference, or something that would encode as a URI ref)

Change of meeting time

Ralph: per my action from last week, the VM Task Force is willing to swap times with us but not next week

RESOLUTION: TF time will move to 1400 UTC starting 24 Jan

Making a Home for the RDF/A Syntax and CURIE Syntax documents

Ralph: unofficial yet, but intention is to ask for a 3-month extension to SWBPD to complete docs. This task force likely will make a case to finish the RDF/A Primer document. This charter extension is a W3C management decision, not AC shouldn't remove pressure on us though

Steven: first HTML WG telecon of 2006 is tomorrow; expect "heads-down" to get final edits done and Last Call published ASAP

Ben: where will the RDFA-syntax document material go?

Steven: expect to cite that document

Ralph: I was under the impression that the RDFA-syntax material would be merged into the XHTML document

Mark: yes, the content will go into the XHTML spec. This takes care of an initial concern that it was important to incorporate the material into the spec proper

Steven: the content will be part of the XHTML2 specification, yes

Ben: the previous XHTML WD had a much shorter version, which wasn't sufficient for an implementor; e.g. Jeremy. Where will the extended prose that deals with the edge cases live?

Steven: with the XHTML2 spec we want to avoid what happened with the HTML 4 spec that a single document tried to address both implementors and authors. We hope to create a normative specification and a separate tutorial document

Ben: do we need to publish RDFA-syntax as a separate document?

Mark: if RDF/A normative is part of XHTML2, then why pull out the RDF/A material as a primer?

Ralph: mostly for sharing the workload. I having a standalone RDFA syntax document creates other worries. Pragmatically, we don't have sufficient resources to maintain 3 separate docs; it would be easier to advance RDF/A as part of XHTML2 for now, split it out later

Mark: I've been thinking recently about an XHTML1 module

Ralph: this TF should complete this design in the context of XHTML2, then once completed, extract it as a module. I am worried about delaying this work too much by trying to generalize it

Steven: XHTML2 is in fact designed as a bunch of new modules. You can't do this in XHTML1 modularization because RDF/A uses features that are not in XHTML1 modularization; e.g. in XHTML1 modularization href is restricted to certain elements so doing RDF/A in XHTML1 requires a revision of XHTML1 modularization. XHTML1 is itself a set of modules but the document of modularization for XHTMl1 is being worked on in the background. Once the new version of modularization is published you would be able to use any module with any of the languages but there are still restrictions on use with XHTML1 for such things as href and attribute name clashes

DanBri: I'd like a public statement of intent to make RDF/A usable with XHTML1. In hallway conversations people are expressing support of RDF/A but don't want to wait for XHTML2 to be deployed

Steven: yes, but still needs a rev to XHTML1

Mark: even XHTML1.1?

Steven: yes, still need a rev even to XHTML1.1

DanBri: would the CDF Working Group be a possible deployment path?

Steven: still have to deal with the attribute clashes between XHTML 1.1 and RDF/A; all the attribute names used by RDF/A exist already in XHTML 1.1Certain elements would end up with two attributes of the same name. There is a common set of attributes that can exist everywhere plus element-specific attributesso some elements would end up with two "href" attributes

Mark: we could recreate the anchor module to remove href

Steve: yes, exactly

Mark: in the 1.1 schemas I demonstrated to the Compound Document people a way to redefine the 'A' module to remove href; don't need to recreate the "A" module

Steven: not sure; I don't think redefinition allows you to remove an attribute from an element

Ben: are we in agreement that as a statement of principle we want RDF/A to be useful with XHTML1?

Steven: every group wants to add their one little thing to XHTML1 so we'd end up with 10 versions of XHTML1 rather than 1 language that has everything. This creates a divergence of markup languages in W3C

<jeremy> +1 to Steven

Steven: and doesn't solve the problem all at once as it ought to

Ben: in any case, making RDF/A work with XHTML1 is not going to happen in the remaining 2-3 months of this Task Force

Ralph: not sure I was comfortable with the blanket high-level statement as there is a lot of work necessary to make that happen, as Steven was pointing out

Jeremy: from the viewpoint of a SemWeb implementation of RDF/A I would expect, for example, that an implemenation like Jena would recognize XHTML documents containing RDF/A natively and could be configured to recognize other documents that contain the RDF/A markup. In practice it's much more important to get the XHTML2 spec right than to worry about legacy documents; each community will deal separately with its user base

<danbri2> (oh sorry; didn't presume to add the rel thing to this meeting... but that we should get it back on agendas... hopefully be back in a few mins)

Mark: in the examples I sent to Ben, most used features that can be done now in XHTML. Should we think about a two-step process; e.g. putting an email address into a page and using rel="foaf:mbox". We could look at what is possible now and later add to the language. rel="foaf:mbox" is like what microformats are trying to do but they have to import all the vocabularies

Ben: it's clear that RDF/A has to function well within XHTML2 first and foremost, so where is syntax specified for XHTML2?

Steven: the XHTML2 syntax is all in schemas. The specification does give the exact content model in each module. There is special notation that indicates when something affects other modules. So each chapter (module) describes its content model and an appendix contains the whole schema; so there'd be a schema module for RDF/A. It may be two modules; one that includes href and another that has the rest

Ben: considering what you expect to have in the XHTML2 spec plus the RDF/A primer, would there be anything missing for an implementor?

Steven: don't think so but an XHTML2 implementor might want more details on the mapping to RDF

Mark: will any of Jeremy's implemenation work contribute to test cases?

Jeremy: only a little. The Jena implementation may have a test suite, though. But my goal is not the same level as represented by the RDF and OWL test cases

Ben: the hole in the specs seems to be in the mapping of RDF/A to RDF

Steven: I would hope that this is part of the XHTML2 document in the minimal form necessary to make it correct and complete

Ben: this (terseness & correctness) is what the RDFA-syntax document was aiming to do. What more does this task force need to do to help the HTML WG?

Steven: I think we have what we need; it's just the mechanics of editing

Ben: I want to make sure that the RDFA-syntax document from which the XHTML editors will be pulling is up-to-date with all our recent decisions

ACTION: Steven inform the task force of a timeline for when a final RDFA-syntax document is needed by the XHTML2 editors [recorded in http://www.w3.org/2006/01/10-swbp-minutes.html#action01]

<Steven> Done!

Ralph: I expect that very shortly our TF focus will switch from looking at RDFA-syntax to looking at XHTML2 editor's draft to be sure our decisions are recorded

Steven: yes

RDF/A Primer

<danbri2> can we put link types draft on the agenda for a future call i'll take care to join?

Ben: one option would be to freeze the current editor's draft as a first WG Note and consider adding examples in a future revision

Mark: I am very conscious of readers' initial impressions of a document. I would not like the current editor's draft to remain as is for several months and be further cited

Ben: the Task Force needs to present a document to the SWBPD WG for review and publishing before the WG charter expires

Ralph: the content of the CURIE specification document is to be merged into the XHTML2 document also, correct?

Steven: yes, that is our expectation

[adjourned]