See also: IRC log
<DanC> HarryH, did you see the agenda edits I made?
<DanC> (if not, I'll go over them by voice; not a big deal)
<DanC> Teleconference minutes for 11 Apr 2007
<HarryH> PROPOSAL: To approve http://lists.w3.org/Archives/Public/public-grddl-wg/2007Apr/0068.html as the true minutes
<HarryH> RESOLVED: http://lists.w3.org/Archives/Public/public-grddl-wg/2007Apr/0068.html are the true minutes for April 11th
<HarryH> PROPOSED: to meet again Wed, 25 April 11:00-0500
<HarryH> Fabien will be the scribe April 25th
<HarryH> RESOLVED: to meet again Wed, 25 April 11:00-0500
<DanC> proposal, basically adopt tests: http://lists.w3.org/Archives/Public/public-grddl-wg/2007Apr/0080.html
<DanC> danc: (don't do much spec text; just tests)
<DanC> jjc: 3 parts to the issue: (a) XHTML, (b) redirects, (c) xml base details
<DanC> Scribe: jjc-scribe
c) xml-base
At LC no normative dependency on XML Base
Issues of conflicts between XPath and XML Base - DanC is confident to sort it otu
Chime: can't do element level base-uri within XPath
(spec editing issue)
DanC: can you change your code to support xml:base (on root element)
No one currently implement xml:base
DanC: leaving aside difficulties of writing spec, how hard to implement
Jeremy: easy
Chime: not hard in 4suite
<HarryH> ACTION: jjc to produce a test-case using xml-base and a relative reference to transform. [recorded in http://www.w3.org/2007/04/18-grddl-wg-minutes.html#action01]
HarryH: does this link to RDFa issue
<Zakim> DanC, you wanted to bring up the stdlib and to
Fabien: no
Library:
Jeremy: profileTransformation should be handled like in transformation
<HarryH> Fabien is noting that RDFa needs to support xml:base.
<FabienG> see http://www.w3.org/2006/07/SWD/RDFa/syntax/#id0x036df498
open whether embeddedRDF should support xml:base
<HarryH> I personally think this would be an argument for eRDF supporting xml:base as well.
<DanC> . ACTION jjc: enhance profileTransformation in our stdlib to handle xml:base
<DanC> . ACTION jjc: test that our stdlib profileTransformation handles xml:base
<DanC> ACTION: jjc to enhance profileTransformation in our stdlib to handle xml:base [recorded in http://www.w3.org/2007/04/18-grddl-wg-minutes.html#action02]
<DanC> ACTION: jjc to test that our stdlib profileTransformation handles xml:base [recorded in http://www.w3.org/2007/04/18-grddl-wg-minutes.html#action03]
<john-l> http://www.w3.org/2006/07/SWD/RDFa/syntax/#id0x03842788
<HarryH> jjc thinks that the test case was broken due to its handling of 3xx..
<DanC> jjc: we're agreed (little a) that redirects should work different from the test suite and implementations
<john-l> http://lists.w3.org/Archives/Public/public-grddl-wg/2007Apr/0066.html
http://lists.w3.org/Archives/Public/public-grddl-wg/2007Apr/0066
Note that if the retrieval was the
result of a redirected request, the last URI used (i.e., that which
resulted in the actual retrieval of the document) is the base URI.
After a redirect GRDDL follows RFC 3986
<DanC> (I'm happy to (a) leave issue-base-param closed, (b) approve the tests as and when they're ready, and (c) cite the tests from the spec. but if somebody wants to make a WG decision today, I don't mind.)
I propose that After a redirect GRDDL follows RFC 3986
Dan summarizes
If you have on root a transformation with "../ aaa.xsl" and if you have an xml:base
then you expect xml:base to work
john-l: also related is http redirects
(sorry john)
<chimezie> I was hoping we PROPOSE to support xml:base on the root and follow RFC 3986 (with the assumption we have no *other* base)
I am happy to second that proposal
<DanC> jeremy, note that the root element could be in an external entity, so the situation isn't as simple as your mail msg suggests.
oh
murray: but we are specifying transforms on XPath node sets, so we don't have uri access
scribing not very accurate
with respect to the base IRI of E
<john-l> Where do we say "resolve URIs based on these other specs" in the GRDDL spec?
GRDDL spec says "with respect to the base IRI of E"
discussion of XML Base and XPath ....
<DanC> (I'm pretty sure XML base doesn't specify conformance of agents)
No conformance section in XML base
Issue about XSLTs
<chimezie> john-l: we say that in the green boxes about grddl:transformation
<HarryH> ACTION: DanC to edit the text of the GRDDL Spec such that a normative dependency on xml:base is in the spec. [recorded in http://www.w3.org/2007/04/18-grddl-wg-minutes.html#action04]
<chimezie> we say: we resolve WRT to the base URI of E (E has no such property)
discussion of XML Base and XPath ...
<FabienG> http://lists.w3.org/Archives/Public/public-grddl-wg/2007Apr/0060.html
Fabien: in xslt 1.0 can't access base URI
<FabienG> http://www.w3.org/2006/07/SWD/RDFa/testsuite/testcases/000004.html
<chimezie> it is up to the application invoking the XSLT to pass in the Base URI of the document entity
<DanC> oh goody... test case. this stuff is frustrating without specifics.
Fabien how to get to base URL in GRDDL
<chimezie> the GRDDL-aware agent is responsible for providing 'default' URIs (both to the transform *and* the parsing of the GRDDL result)
DanC teh xslt will output rdf:resource="photo1.jpg"
and the right thing will happen
<chimezie> The XSLT will *not* try to attempt to resolve photo1.jpg against the base URI of the document entity, because it won't interpret it as a relative URI
jjc: teh base URI of the original document is used as the base URI of the RDF/XML
Fabien: yes that does solve my problem
<Zakim> DanC, you wanted to ask if SWBD gave a date re "SWBD will give us comments or explicit no comment."
HarryH: can we take test cases to LC and spec to CR/PR: CG said no
everything to PR at the same time, hence likely delayed PR to after www 2007
<DanC> (the CG is basically just reading the process document)
<FabienG> doesn't know if we should talk about RDFa WG issue on hGRDDL requiring our cooperation http://lists.w3.org/Archives/Public/public-grddl-wg/2007Apr/0053.html as mentioned in the agenda.
<DanC> routes that make sense to me: (1) CR-spec + LC-tests, (2) skip CR and go to PR later
We need to have LC on test cases, hence three week LC period, and then PR on everything
jjc: primer either needs to be ready for LC with test cases, or not rec track
chime: requests review of editors draft of Test cases
danc: no one else available
harry can give a light read of tc
jjc: a few tests to add, a few
bugs to fix in existing tests
... i think we agree where the bugs are
May 2nd is last date on which we can agree on CR/LC
Danc: no too late
<HarryH> DanC would prefer April 25th for CR.
HarryH: 25th Apr is then the only possibility
DanC does not have a preference
DanC and HarryH have other commitments and things
What's the rush?
<DanC> (it's been dotting i's and crossing t's, for me, since the WG convened.)
Murray: advocated taking things a bit slower, particularly given that the bulk of work is done
<HarryH> ACTION: HarryH to write CR report request by early next week and Jeremy to review. [recorded in http://www.w3.org/2007/04/18-grddl-wg-minutes.html#action05]
<HarryH> Jeremy? Happy with your EARL?
<HarryH> ACTION: HarryH to run EARL-> HTML table generator and work on CR/PR transition request [CONTINUES] (to give to jjc by next week if not done) [recorded in http://www.w3.org/2007/04/18-grddl-wg-minutes.html#action06]
<DanC> proto-impl report, with earl foo: http://www.w3.org/2001/sw/grddl-wg/td/test_results
samsung will be invited to waive their rights
<chimezie> http://www.w3.org/2001/sw/grddl-wg/td/grddl-tests#rdfXMLDoc
<DanC> Harry notes that unless recorded otherwise, actions in the agenda continue.
<HarryH> PROPOSAL: to approve http://www.w3.org/2001/sw/grddl-wg/td/grddl-tests#rdfXMLDoc as a valid test caes
<john-l> application/vnd.mozilla.xul+xml
<chimezie> *phew*
<DanC> "Get RDF from a spreadsheet"
jjc: no that's mozilla it silently converts application/rdf+xml to that one
<john-l> "RDF/XML document"
<DanC> no, not "Get RDF from a spreadsheet". never mind. I was confused.
<HarryH> RESOLVED: http://www.w3.org/2001/sw/grddl-wg/td/grddl-tests#rdfXMLDoc approved as a valid test case
<chimezie> http://www.w3.org/2001/sw/grddl-wg/td/grddl-tests#xhtmlWithGrddlProfile
<chimezie> "One transform linked from the head of a document with only the GRDDL profile"
<HarryH> PROPOSAL: to approve http://www.w3.org/2001/sw/grddl-wg/td/grddl-tests#xhtmlWithGrddlProfile as a valid test cae
Dan asks about acknowledgement to Dom, who authored test
<chimezie> ACTION chime to update dom in ack
<chimezie> ACTION: chime to update dom in ack [recorded in http://www.w3.org/2007/04/18-grddl-wg-minutes.html#action07]
<HarryH> RESOLVED: http://www.w3.org/2001/sw/grddl-wg/td/grddl-tests#xhtmlWithGrddlProfile approved as a valid test case
<chimezie> http://www.w3.org/2001/sw/grddl-wg/td/grddl-tests#xhtmlWithGrddlTransformationInBody
<chimezie> "One transform linked from the body of a document with only the GRDDL profile"
<HarryH> PROPOSAL: to approve http://www.w3.org/2001/sw/grddl-wg/td/grddl-tests#xhtmlWithGrddlTransformationInBody as a valid test case
<HarryH> RESOLVED: http://www.w3.org/2001/sw/grddl-wg/td/grddl-tests#xhtmlWithGrddlTransformationInBody is a valid test case
<chimezie> http://www.w3.org/2001/sw/grddl-wg/td/grddl-tests#xmlWithGrddlAttributeAndNonXMLNamespaceDocument
<chimezie> http://www.w3.org/2001/sw/grddl-wg/td/grddl-tests#sq2
<DanC> (I'm only encouraging a batch if there's nothing to discuss, as far as you know)
<chimezie> http://www.w3.org/2001/sw/grddl-wg/td/grddl-tests#two-transforms
<chimezie> those 3 for now
jjc: seconds all three
<HarryH> RESOLVED: #xmlWithGrddlAttributeAndNonXMLNamespaceDocument, #sq2, and #two-transforms are all valid test cases, with a base IRI of http://www.w3.org/2001/sw/grddl-wg/td/grddl-tests
<chimezie> http://www.w3.org/2001/sw/grddl-wg/td/grddl-tests#four-transforms
<chimezie> http://www.w3.org/2001/sw/grddl-wg/td/grddl-tests#pf-pf-pf-ns
<chimezie> http://www.w3.org/2001/sw/grddl-wg/td/grddl-tests#ns-ns-pf-pf
<chimezie> http://www.w3.org/2001/sw/grddl-wg/td/grddl-tests#ns-pf-ns-pf-ns
DanC concurs on this one
<chimezie> http://www.w3.org/2001/sw/grddl-wg/td/grddl-tests#html-and-transformation-attr
(one = the last four)
<HarryH> RESOLVED: #four-transforms, #pf-pf-pf-ns, #ns-ns-pf-pf, #ns-pf-ns-pf-ns are all valid test cases with a base URI of http://www.w3.org/2001/sw/grddl-wg/td/grddl-tests
<chimezie> http://www.w3.org/2001/sw/grddl-wg/td/grddl-tests#grddlonrdf-xmlmediatype
<DanC> (I concurred after I heard jjc and chime report succesful experience with the 4 including #ns-ns-pf-pf . I'd feel better if I had run my cwm/n3 impl on them, but I'm OK with having that happen another day, if ever.)
<DanC> chime withdraws #xhtmlWithGrddlEnabledProfileAndADisabledInBodyTransform
jjc was grddlonrdf-xmlmediatype previously known as grddlonrdf-xmlmediatype3
chime: yes
<chimezie> http://www.w3.org/2001/sw/grddl-wg/td/grddl-tests#unknown-media-type
jjc: is unknown-media-type informative
chime: yes
<HarryH> PROPOSE: #grddlonrdf-xmlmediatype and #unknown-media-type are valid test cases with a base URI of http://www.w3.org/2001/sw/grddl-wg/td/grddl-tests
<HarryH> PROPOSE: #grddlonrdf-xmlmediatype is a normative test case and #unknown-media-type is an informative test cases and both have a base URI of http://www.w3.org/2001/sw/grddl-wg/td/grddl-tests
<DanC> (I'm uncomfortable with "informative" but I undestand better now.)
Dan was does informative mean?
jjc: like we have no normative
dependency on turtle
... and we also have no norm dependency on the unknown media
type
... I second
<HarryH> RESOLVED: #grddlonrdf-xmlmediatype is a normative test case and #unknown-media-type is an informative test cases and both have a base URI of http://www.w3.org/2001/sw/grddl-wg/td/grddl-tests
<chimezie> http://www.w3.org/2001/sw/grddl-wg/td/grddl-tests#hl7-to-owl
<DanC> chime: did you (jeremy) pass #hl7-to-owl? I recall some discussion of the urn
<chimezie> hl7-rim-to-pomr.xslt
jjc: my system had a bug, but is now fixed
<HarryH> http://www.w3.org/2001/sw/grddl-wg/td/grddl-tests#hl7-to-owl is a valid test case.
<HarryH> RESOLVED: http://www.w3.org/2001/sw/grddl-wg/td/grddl-tests#hl7-to-owl is a valid test case.
<FabienG> Pointer to hGRDDL http://www.w3.org/2006/07/SWD/wiki/RDFaGRDDL
<chimezie> hGRDDL would output 'application/xhtml+xml' mimetype
<chimezie> is that sufficient for a mime-type-driven follow-your-nose GRDDL result parsing?
<HarryH> ACTION: Fabien to take action on testcase for microformat+XHTML->RDFa (i.e. hGRDDL). [recorded in http://www.w3.org/2007/04/18-grddl-wg-minutes.html#action08]
<DanC> I'm interested in dicsussion of # mechanical rules (GRDDL spec editorial)
Fabien: can merge outputs of grddl transforms
<chimezie> .. interesting ..
Fabien: can merge outputs from different microformats
<Zakim> jjc-scribe, you wanted to talk about goals for next week and to propose extension
agreed to extend to the hour
jjc: what do we need to have done by next weel?
<chimezie> i think a goal of doing all we can to ensure a smooth transition to an XML Base norm. dependency would be very useful
jjc: need to freeze docs at least 24hrs before meeting
<HarryH> Editors can send out an e-mail if they do or do not feel ready to freeze documents?
jjc: chime what's the deadline for test cases
chime: the weekend
<DanC> I'll do what I can to inform you in time for T-24 agenda release, harry
<chimezie> lol
jjc: I would prefer them to move, but don't demand it
john-l: I like them
harry: others find them confusing
chime: I like them where they are
danc: if I had more time, I would move them, but ....
<DanC> "The formal specification of this markup is given below. An informative mechanical version of each rule is given with the premise and the conclusion written as SPARQL graph patterns[SPARQL]. See the Mechanical Rules appendix for namespace prefix bindings and further explanation."
<scribe> ACTION: jjc to think of some text to soften the above [recorded in http://www.w3.org/2007/04/18-grddl-wg-minutes.html#action09]
discussion of microformats and grddl
text about scraping was removed from use cases note
http://www.w3.org/TR/2007/NOTE-grddl-scenarios-20070406/#html_tidy_use_case
<HarryH> http://www.w3.org/TR/grddl-scenarios/#html_tidy_use_case
Although tidying of source documents can be part of a pragmatic approach to gathering data, the consumer of the RDF can only trust GRDDL transformations when they have been explicitly licensed by the author of the documents.
<DanC> ----------------------------
<DanC> revision 1.65
<DanC> date: 2006/12/25 06:20:19; author: hhalpin; state: Exp; lines: +15 -11
<DanC> removed word scraping, added caveat for non-XHTML GRDDL
<DanC> ----------------------------
<DanC> (I stipulate that it's procedurally strange. so?)
what will the audience at www 2007 say about grddl?
adjourn
rssagent, please publish the minutes