See also: IRC log
Steven: the latest
editor's draft is built nightly, not yet uploaded to W3C
site
... please take a look and give us feedback
... the area we're having problems with still is blank
nodes
... we want to create anonymous nodes and give them names
independent of URLs
... for this one little edge case it seems a new naming and
referencing mechanism needs to be created
... e.g. aboutblank and hrefblank
... and we need a way to give a (blank) name to an
element
... in RDF/A there was a proposed XPointer framework for
referring to blank nodes and a naming mechanism
... the problem with this XPointer framework solution was that
we'd have to create a spec for it, as it's not defined
anywhere
Mark: we could simply define that
XPointer framework as in the RDF/A draft -- we don't have to go
all the way back to XPointer
... the framework makes use of the XPointer architecture
... we could define this in our spec, don't need to go back to
XPointer WG
Jeremy: I didn't like the RDF/A
XPointer solution
... I found it "ugly"
... one nodeID attribute is sufficient; no need for extra
hrefblank
... what's hard is providing text for an HTML author who is not
RDF-aware to motivate this nodeID attribute
Steven: you're suggesting a nodeID attribute that simultaneously defines and references?
Jeremy: yes
... in the [October] RDF/A proposal, if there was not a subject
in the triple then it was inherited from the context
... if nodeID expresses the subject it's possible to default
the object and vice-versa
... while clunky, this achieves the goal of being able to
express all of RDF without making things too complicated for an
HTML author
... [the spec is] affected by the complexity of the mapping
rules from the XHTML attributes to RDF
... I felt the [October] RDF/A mapping rules were a bit too
complex
Mark: I came up with the XPointer
solution because I'm pretty sure we do need two
attributes
... RDF/XML does not need two attributes due to its striping
syntax
... in the XHTML mechanism we've tried to make it possible to
write a complete triple in a single element
... Jeremy's suggested defaulting mechanism may work but I'm
pretty sure there are cases where you still need two
attributes
... it's also an aesthetic problem for bnodes to have URIs that
are not supposed to be considered as URIs
Ben: is the bnode section in this editor's draft yet?
Steven: no. the document I cited
has RDF/A transformed into XHTML2 style except for the bnode
material
... we've been struggling with finding some mechanism for
bnodes that is more aesthetic and that is explicable to HTML
authors without referring to the term "bnode"
Jeremy: I agree with that objective
Mark: why would an author use
anything other than id="A" and id="B" to refer to anonymous
nodes?
... why should I [as a document author] be worried about any
outside use of those names?
... e.g. I use those names locally, but why should I prevent
them from use outside?
... I think this is a concern more to RDF folk than to an HTML
author
Ben: let's move this discussion to e-mail
Jeremy: at the March face-to-face, Steven suggested that XHTML2 would go to Last Call without addressing this particular issue
ACTION: Ben to move bnode discussion to email list. [recorded in http://www.w3.org/2005/04/05-swbp-minutes.html#action01]
Steven: we don't expect any given
XHTML spec to have addressed all open issues; we simply have to
freeze a document and count remaining open issues as Last Call
issues
... if we didn't do this we'd never get a document out the
door; people are always reading new drafts and bringing new
issues
Ralph: are the words in the editors' draft now fairly stable? would a careful review be a waste of time?
Steven: not a waste of time
... people are fixing schema errors now
... but up to publication there is always the chance of
changes
ACTION: Steven to send email about latest draft of RDF/A included in XHTML 2 [recorded in http://www.w3.org/2005/04/05-swbp-minutes.html#action02]
Steven: the editors' draft cited above has dealt with all the issues from the HTML WG F2F
ACTION: BenA to Examples in RDF/XHTML and RDF/XML - use cases [recorded in http://www.w3.org/2005/04/05-swbp-minutes.html#action03] [reworded below]
Ben: any specific examples you'd like to see, or just use cases?
Steven: use cases good
Mark: it is OK to give RDF/XML
examples and we'll translate them
... most of the larger examples I've done have been based on
RSS and FOAF. These might not express everything
ACTION: BenA Provide RDF/XML examples and english description to Steven and Mark (use cases) [recorded in http://www.w3.org/2005/04/05-swbp-minutes.html#action04] [CONTINUES]
ACTION: Tom Baker and Gavin to get feedback about use of RDF in XHTML in their respective communities [recorded in http://www.w3.org/2005/04/05-swbp-minutes.html#action05] [CONTINUES]
ACTION: DanBri RDF schema for new XHTML2 namespace elements [recorded in http://www.w3.org/2005/04/05-swbp-minutes.html#action06] [CONTINUES]
Ben: let's plan to meet weekly, even if it's a short meeting
next meeting: 12 April, 1400 UTC
Ben: discussing f2f feedback; do we need to address a pre-XHTML2 solution?
Jeremy: I'll take an action to get HP feedback
ACTION: Jeremy to ask HP about need for pre-XHTML 2 solution to RDF-in-HTML problem [recorded in http://www.w3.org/2005/04/05-swbp-minutes.html#action07]
Ben: Dan and Dom did a new editor's draft for GRDDL
Ralph: I read the new draft but didn't look at the diffs
Ben: what I gleaned from the
email is that typos were fixed
... and specifying what happens if more than one transform is
given within a document
Ralph: Dave Beckett
announced an implementation
... Dave asked if multiple tranformations yield a single
graph
... Dom answered in the affirmative
Ben: what communities do we need to approach?
ACTION: Ben to ask Tom, Gavin and CC about opinion on GRDDL and pre-XHTML2 [recorded in http://www.w3.org/2005/04/05-swbp-minutes.html#action08]
Ralph: Creative Commons and
Dublin Core are still in the top of our list.
... we could grow the list but at least we need those two
responses
<Zakim> Jeremy, you wanted to summarize changelog and to discuss possible questions to users
Jeremy: I see only editorial
changes in the GRDDL
changelog
... should I be asking HP "Is GRDDL useful" or "Do we need a
solution to embedding RDF before XHTML2"?
Ben: I think the latter; do we need a solution before XHTML2 and if so, is GRDDL a good way to do this?
Jeremy: the vast majority of HTML
on the web is ill-formed, not XML
... if we're seeking user feedback, it's up to them to define
the scope
Ben: I think it's good to cast a
wide net to find out user expectations
... this task force may not last long enough to provide every
solution, but please ask
Ralph: Dave Beckett's GRDDL
implementation partly answers some questions that Ivan Herman
raised in a hallway conversation at the March F2F
... note that the GRDDL spec says transformations SHOULD be
XSLT, not MUST
... so implementations have to be careful