See also: IRC log, previous 2007-06-21
Michael: @about in HEAD is causing 2 of the test cases to not validate on new W3C validator; cases 12 and 13
Shane: @about is not permitted on HEAD
... there's a collection of attributes called COMMON, which includes
@about
... but HEAD does not accept COMMON
... this may be more of an XHTML question
all actions still PENDING
Michael: we should review the XHTML1.1 test cases today
<mhausenblas> wiki discuss
RDFaTC (last edited 2007-06-28 12:19:20 by MichaelHausenblas)
<Steven> I get 404 on the actual tests
Michael: I removed a couple of cases because they were out of date w.r.t. current syntax
Michael: the -> manifest file describes the structure
ACTION: Michael create a .htaccess hack to make the test case identifiers resolve to something useful [recorded in http://www.w3.org/2007/06/28-rdfa-minutes.html#action01]
Michael: let's go through to decide which are
valid
... do view source on the .xhtml and .sparql
<mhausenblas> human-readable view of manifest
<mhausenblas> ManifestFiles
<Ralph> [trying to load the manifest.rdf into Tabulator ]
Shane: we're talking about XHTML1.1 with RDFa
annotation
... focussing for the moment on META and LINK, Mark and I have extended the
content model for META and LINK
... we originally did this in the context of XHTML2
... is it right to be pulling this back into XHTML1.1 ?
Mark: it was right for XHTML2
... it has since emerged that browsers move LINK and META back into HEAD
... so we have an issue to consider removing that functionality
Shane: the issue has been raised about META and
LINK in BODY
... but no issue yet on content in META and LINK
... so it occurs to me now that these test cases are probably wrong [for
XHTML1.1]
Mark: this comes back to the question of
changing the host language
... in terms of deployed browsers, it's difficult to permit content in META
and LINK
... <META> and </META> are two different elements in [current]
parsers
... so nothing after <META> is not currently interpreted as a child of
META
Michael: I believe TCs 2 and 3 should not be
considered right now for XHTML+RDFa
... so the result of the SPARQL query on TC 1 should be TRUE
Shane: all tests should be structured so TRUE result means 'pass'
Mark: what did we decide about language?
Michael: see Section 3 of Test Suite overview
Mark: best practice in XHTML is to include xml:lang
Shane: on the HTML element?
Mark: yes, but need to check that this is current best practice
Steven: yes, normally the HTML element should have @xml:lang but I'm not sure what is best for our test cases; do we want a lang triple?
Ralph: so we do not expect that there is more than one triple that MUST be generated in TC1?
Michael: yes
RESOLUTION: TC1 accepted
Michael: 2 through 4 should be removed until syntax questions are settled
RESOLUTION: TC2, TC3, and TC4 to be put 'on hold'
Michael: I suggest removing TC5 as well
RESOLUTION: TC5 to be removed
<Ralph> [yay! manifest.rdf does load in Tabulator. The Purpose property seems to be the description I'm looking for]
<ShaneM> There are assertions in here: the manifest
Mark: we shouldn't drop an @xml:base test -- so we should revise TC4 rather than remove it
<mhausenblas> http://validator-test.w3.org report
Shane: it's ok to put @xml:base on SPAN but the test then needs to put something in the content to test
Michael: but the markup doesn't currently validate
ACTION: Shane investigate the @xml:base validation issue [recorded in http://www.w3.org/2007/06/28-rdfa-minutes.html#action02]
Shane: @xml:base should work everywhere; the DTDs may be broken
Michael: I don't want to propose any test cases that the W3C validator rejects
Shane: but this case seems to be a bug in the validator, which I'll look into
Michael: call test case 4 "on hold"
Mark: similarly, TC2 and TC3 have no particular reason to be in BODY; the important parts of those tests could just as well be done in HEAD
<Steven> http://www.w3.org/TR/xhtml-modularization/abstract_modules.html#s_structuremodule
ACTION: Shane to correct DTD to permit RDFa attributes on the head element. [recorded in http://www.w3.org/2007/06/28-rdfa-minutes.html#action03]
Michael: so TC2 and TC3 should be changed to
"on hold" and I will move the LINK and META to the HEAD
... TC6 is about @rel and @rev
Steven: was concerned for a moment about foaf:img but looks fine
RESOLUTION: TC6 accepted
Mark: can we get rid of excess namespaces at the top?
Ralph: yeah, I think stripping the test cases down to the essentials is better
Steven: I'm happy with TC7
Shane: I agree that test cases should be minimal; excess stuff could be a perturbation
RESOLUTION: TC7 ok
TC8 - empty string @about
Ralph: the triple would be the same if @about were missing
Steven: that's worth testing also
Ralph: I'd suggest a separate test case that omits @about
RESOLUTION: TC8 ok
ACTION: Michael add a test case that test for omitted @about [recorded in http://www.w3.org/2007/06/28-rdfa-minutes.html#action04]
TC9 - @rev
Mark: might want to change the TC order; e.g. test @rev, @rel independently then test the combination
Shane: good test cases are independent, which includes order
Mark: given that we're using example.org I
think it's worthwile adding fragment identifiers
... so that these aren't information resources
RESOLUTION: TC9 ok, with #frag added
TC10 - @rel, @rev, @href
RESOLUTION: TC10 ok, with #frag added
TC11 - XMLLiteral
Shane: the SPARQL doesn't escape '<' and '>' -- does it need to?
Michael: that's the way RDF defines XMLLiteral
Ralph: the SPARQL expression is not itself encoded in XML so the special characters do not need to be escaped there
RESOLUTION: TC11 ok
TC12 - @xml:lang
RESOLUTION: TC12 ok
TC13 - @xml:lang inheritance
Michael: same issue with @about
RESOLUTION: TC13 ok
TC14 - @datatype, xsd:integer
RESOLUTION: TC14 ok
TC15 - Blank node via meta and link
RESOLUTION: TC15 ok
Steven: does 15 really produce a blank node?
Michael: the title is misleading
ACTION: Michael correct TC15 title [recorded in http://www.w3.org/2007/06/28-rdfa-minutes.html#action05]
TC16 - Blank node, explicit
Michael: propose to put TC16 on hold
Mark: the test does what it claims, the question is about CURIEs
Michael: same applies to TC17
RESOLUTION: TC16, TC17 on hold pending CURIE decision
TC18 - @rel for predicate
Shane: what does this do differently?
Mark: this tests @rel alone
RESOLUTION: TC18 ok
[adjourned]
<mhausenblas> Thank you all for the hard work!!!
<Ralph> [I think you did most of the truly hard work, Michael!]
<Ralph> [or, at least, the tedious work]
<mhausenblas> well elias and wing did - i just had the pleasure to present it ...
<mhausenblas> more or less ;)
ACTION: [DONE] Ben send email to SWD WG list summarizing RDFa document status [recorded in http://www.w3.org/2007/05/31-rdfa-minutes.html#action02]
-> An Update on RDFa Schedule and Docs
ACTION: [DONE] Ben start a list of RDF/XML features that are not supported by RDFa [recorded in http://www.w3.org/2007/01/23-swd-minutes.html#action01]
-> RDFCoverage
ACTION: [DONE] Wing add a property to the test case schema for tracking origin and approval of an individual test [recorded in http://www.w3.org/2007/03/05-rdfa-minutes.html#action11]
<Ralph> actually, it's fun to look at the manifest in tabulator
<Ralph> I'll think about what may make it even more fun :)
<mhausenblas> I'll have a look (but zitgist is alos nice :)