See also: IRC log, previous 2007-08-02
<RalphS> Thanks, Michael, for proposing an agenda
ACTION: Ben to recontact implementors Elias, MarkB, triplr and post their implementations to http://esw.w3.org/topic/RDFa#Implementations [recorded in http://www.w3.org/2007/08/02-rdfa-minutes.html#action09] [CONTINUES]
ACTION: Ben to work through xml:lang issue with Ivan [recorded in http://www.w3.org/2007/08/02-rdfa-minutes.html#action07] [CONTINUES]
ACTION: [DONE] Michael test how <SPAN /> behaves in Internet Explorer [recorded in http://www.w3.org/2007/08/02-rdfa-minutes.html#action06]
-> span test
ACTION: [DONE] Michael to email individual emails to resolve TCs 29-33 (TC 28 to be resolved post xml:lang discussion) [recorded in http://www.w3.org/2007/08/02-rdfa-minutes.html#action08]
ACTION: Ben to look into Science Commons use case [recorded in http://www.w3.org/2006/12/11-htmltf-minutes.html#action04] [CONTINUES]
ACTION: Ben to sum up @href/@resource everywhere proposal [recorded in http://www.w3.org/2007/07/12-rdfa-minutes.html#action10] [CONTINUES]
ACTION: Elias to send email to list with use case from IBM [recorded in http://www.w3.org/2006/12/04-htmltf-minutes.html#action10] [CONTINUES]
ACTION: MarkB to work rdf:label back into RDFa syntax when using @content after October [recorded in http://www.w3.org/2007/07/19-rdfa-minutes.html#action08] [POSTPONED]
<mhausenblas> RDFa TC
Michael: test 28 is still on hold, pending Ben's action
-- test 29; meta content --
Michael: Ben suggested changing the title
Michael: should the final space be dropped?
Ralph: this has to do with the default for preserving whitespace in XML parser, right?
Shane: yes, and XHTML requires that the parser always preserves whitespace
Ralph: in that case, then I believe the triple does need to retain the whitespace
resolved: test 29 approved, after correcting the result to include the whitespace and updating the title per Ben's suggestion
-- test 30, omitted @about --
-> test 30
Michael: I made the Creative Commons correction requested by Ben
resolved: test 30 approved
Ralph: Ivan commented that it could be useful
to have @xml:base in our tests so the test can be run off-line
... I feel that adding @xml:base changes the tests and it's easy enough for a
tester to modify the subject URI if they need to do so
Mark: perhaps we could add a $BASE variable to the results so it's clear what would need to be replaced
Michael: I say that as we're testing the
Semantic Web the tests should be run on-line
... the test should work on the Net
Shane: I brought this up in the XHTML2 WG
meeting again yesterday
... my belief is that @xml:base will not be added to Modularization 1
... because the other languages that depend on Modularization 1 do not define
the semantics of @xml:base
... if this group does want @xml:base support then it's trivial to add
... this decision would have far-reaching ramifications, however
... we'd be saying that browsers would have to support it
... current browsers don't support it
Mark: let's postpone a decision on @xml:base
ACTION: Michael add @xml:base issue to tracker [recorded in http://www.w3.org/2007/08/09-rdfa-minutes.html#action09]
-- test 31: simple @resource --
-> test 31
Mark: I'd suggest that we not use A in this test; use DIV or SPAN instead
<ShaneM> I agree
Mark: in this particular case, a SPAN around
the ISBN would be nicer
... why not put @about on the P rather than the DIV?
Michael: I think the validator wouldn't let me omit the P
Shane: it won't validate without the P but it
would validate without the DIV
... where's the 'wtw' anchor?
Mark: there doesn't need to be one
Shane: it's an unresolvable fragment id. I don't care if you guys don't care
Mark: it's ok but it does look odd. We could use an Amazon URI instead.
Shane: if this simple minimalist test is OK in
your eyes, that's fine
... this test is speaking about something I can't find
Ralph: if there were an anchor then there is an ambiguity about whether the subject is a book or an HTML fragment
Mark: think of it as a primary key; it just happens to look like a URI
<mhausenblas> LinkedDataTutorial
Shane: the document would fail W3C link
checking
... but maybe that's OK
resolved: test 31 accepted, removing DIV and replacing A with SPAN
-- test 32; @resource overrides @href --
-> test 32
Michael: similar to test 31 except we also have @href and @resource takes precedence
<ShaneM> Again, we could get rid of the DIV and move the about to the P...
<ShaneM> Keep the A
Michael: we need A in this case to illustrate the @href
Mark: I'd prefer that @href point to an information resource
resolved: test 32 accepted, removing DIV and replacing the @href value with an information resource
-- test 33; simple chaining test with bNode --
-> test 33
Michael: test 33 was split out from test 25
Ralph: subject URI still says '25'
Shane: general editorial comment about all
these tests;
... I believe we had discussed dropping the unused xmlns: declarations at the
top of the test cases
Michael: yes, we agreed and didn't record an action
ACTION: Michael remove excess xmlns: declarations from test cases [recorded in http://www.w3.org/2007/08/09-rdfa-minutes.html#action10]
resolved: test 33 accepted, with corrected subject
Michael: regarding test 9
... we approved this in error; the sparql query was wrong
-> test 9
scribe: what should we do process-wise if someone identifies an error in a previously approved test?
Michael: I fixed the error in the sparql query (reversing Person1 and Person2)
resolved: test 9 re-approved, as corrected
Ralph: I think it's good to have on record when we correct a test
Shane: we can be fairly loose now but once
we're in CR then we'll need to version the test suite
... version N of the test suite should only be changed by declaring a test
invalid, not adding or removing tests
... as far as 'editorial' changes, we should rely on Michael's judgement
Michael: I promise to report to the TF whenever I change something
Shane: I think there's only about 3 changes needed. I just need someone to summarize and I'll update the DTD
<ShaneM> meta and link are not everywhere....
<ShaneM> add instanceof and resource
Michael: I suggest we wait to change the DTD until we've resolved [the name of] @instanceof
<markbirbeck> Also, @href is not everywhere.
Michael: also @xml:lang
Shane: two different issues. The good news is
that the new W3C validator relies on the DTD that I control
... the second issue is to get the DTD into the pull-down list, which
probably won't happen until RDFa is at least CR
Ralph: I'd like to be reminding people that we've switched from @class to @instanceof asap
Shane: unfortunately, the validator won't help with that at all, as @class remains DTD-valid
Mark: people are already adopting idioms for RDFa; e.g. <EVENT> ...
<mhausenblas> W3C HTML+RDFa validator
Shane: best we can do is announce the change very publicly and loudly
Michael: the xml:lang decision is still a big
open thing for at least one implementor
... we'll need Ben's action to close that
Shane: this issue comes because we're using
XHTML1.1 strict which doesn't have @lang, just @xml:lang ?
... we deprecated @lang in XHTML 1.0
... the Task Force could decide to add @lang back, but I would not recommend
this
Mark: we should be leveraging XML and XHTML,
not adding too much
... the idea of RDFa being a layer on an existing language, with part of the
definition being the interpreation of existing attributes in the base
language and the other part of the definition being the attributes that we
add
Michael: the test cases could just say "it's XHTML1.1 now"
Mark: or make two sets of tests
Shane: I fully intend to produce a DTD for
HTML+RDFa
... I've already done it and I intend to maintain it
<mhausenblas> see also -> RDFaInHTML
Shane: in the context of HTML4 we do need to
talk about @lang
... as we're inheriting that environment
Michael: I wouldn't like to see that we define @lang one way for HTML4 and a different way for XHTML1
Mark: but host languages are different, so we can't say the interpretation is the same for all host languages
Michael: could we ignore @lang?
Mark: we do talk about typed literals and string literals
Michael: how much is @lang used?
Mark: we don't want to repeat the long
discussion we had over typed literals
... and we don't want to change the host language too much
Shane: [only] XHTML1.0 supports both @lang and @xml:lang
Mark: we could specify how we interpret each
attribute if it is present in the host language
... but not require the host language to provide all attributes
<mhausenblas> RDFa in HTML Overview
Mark: may need to specify this abstractly, such
as Infoset does
... we're talking about a set of processing rules
Mark: e.g. things like relative paths also are
affected by the host language
... we should be honoring the rules of the host language
... it's a matter of interpretation
ACTION: Mark summarize in mail host language-dependent issues such as @lang and @xml:lang [recorded in http://www.w3.org/2007/08/09-rdfa-minutes.html#action11]
Shane: the RDFa Syntax spec should document the
core RDFa thing
... and XHTML+RDFa can be another document that describes how RDFa extends
XHTML, including any additional interpretation rules
... same for HTML+RDFa
Michael: the XHTML+RDFa document(s) could also be the place to put CURIEs
Shane: I think CURIEs need to go into the Syntax document, as the CURIE interpretation needs to apply to all of our new attributes
[adjourned]