See also: IRC log, previous 2007-01-29
<RalphS> Previous: 2007-01-29 http://www.w3.org/2007/01/29-htmltf-minutes.html
<Steven> gosh, look at the time :-)
<michaelhausenblas> ScribeNick: mhausenblas
<RalphS> Scribe: Michael
<RalphS> scribenick: michaelhausenblas
Steven: Spoke at W3C eGov
meeting
... policy oriented people
... Ivan spoke about RDFa as well
<RalphS> European W3C Symposium on eGovernment
Steven: In Sep 2006 we planned to
submit to dev track
... now a new deadline: Feb 12
... I'm planning to participate
... in Banff/Canada
<Steven> My talk: http://www.w3.org/2007/Talks/02-01-steven-egov/
Ben: There are a lot of things to point out
<benadida> ex: http://ben.adida.net/research/
<Steven> Ivan's talk: http://www.w3.org/2007/Talks/0202-Gijon-IH/
<RalphS> Simile Exhibit
<scribe> ACTION: Ben to contemplate about the Dev Track [recorded in http://www.w3.org/2007/02/05-htmltf-minutes.html#action01]
<benadida> quick point: Exhibit is not my project, I take no credit for it :)
Steven: machine tags at flickr
<RalphS> Machine Tags [Karl Dubost]
Steven: namespace are there but not yet URIs
Welcome to the Semantic Web, flickr
Steven: now, a window of opportunity is open
Ben: We should talk to the flickr guys and invite them to go for RDFa
<scribe> ACTION: Ben to get in contact with flickr w.r.t. machine tags [recorded in http://www.w3.org/2007/02/05-htmltf-minutes.html#action02]
http://www.w3.org/2007/01/29-htmltf-minutes.html#ActionSummary
<scribe> ACTION: Ben to put Michael and Wing in contact for test case production [recorded in http://www.w3.org/2007/01/29-htmltf-minutes.html#action17] [DONE]
<scribe> ACTION: Mark to put M12N schema up somewhere, and tell us where [recorded in http://www.w3.org/2007/01/29-htmltf-minutes.html#action01] [CONTINUES]
<Steven> Devtrack deadline is Feb 16th according to website
<scribe> ACTION: Mark write examples/tests of striping support [recorded in http://www.w3.org/2006/12/04-htmltf-minutes.html#action07] [DROPPED]
<scribe> ACTION: Steven to put together sample XHTML2 doc with all mime type, etc. [recorded in http://www.w3.org/2006/12/04-htmltf-minutes.html#action09] [CONTINUES]
<scribe> ACTION: 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] [CONTINUES]
<Steven> http://www2007.org/cfp-Developers.php
<scribe> ACTION: Ben to get the docs in good shape for next week [recorded in http://www.w3.org/2007/01/23-swd-minutes.html#action04] [DONE]
<scribe> ACTION: Ben to look into Science Commons use case [recorded in http://www.w3.org/2006/12/11-htmltf-minutes.html#action04] [CONTINUES]
<scribe> ACTION: Ben to update issues list with the @CLASS overload problem [recorded in http://www.w3.org/2007/01/23-swd-minutes.html#action08] [DONE]
<scribe> ACTION: Ben to update RDFa schedule on wiki to aim for last call on June 1 [recorded in http://www.w3.org/2007/01/23-swd-minutes.html#action09] [CONTINUES]
<RalphS> new (tracker) issues list
<RalphS> Ralph: I have a patch written by a summer student to get tracker to output RDF
<scribe> ACTION: Ben to write down the relation between GRDDL and RDFa [recorded in http://www.w3.org/2007/01/23-swd-minutes.html#action02] [CONTINUES]
<scribe> 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]
<scribe> ACTION: Mark produce more examples of applicability of n-ary relations from IPTC documents [recorded in http://www.w3.org/2006/12/04-htmltf-minutes.html#action08] [CONTINUES]
<scribe> ACTION: Mark produce RDFa XHTML 1.1 module and spec [recorded in http://www.w3.org/2006/12/18-htmltf-minutes.html#action06] [CONTINUES]
<scribe> ACTION: Michael continue work on the FAQ template that Elias started [recorded in http://www.w3.org/2007/01/08-htmltf-minutes.html#action10] [CONTINUES]
<scribe> ACTION: Michael transfer issue data to SWD tracker [recorded in http://www.w3.org/2007/01/08-htmltf-minutes.html#action03] [CONTINUES]
<RalphS> Michael: I wanted to transfer the closed issues
Michael will make a single issue for all closed issues
<RalphS> Ralph: one closed issue pointing to the HTML page would suffice for me
Mark: I have been in contact with
Sean
... will try to push work forward
Ben: This is a cruicial part of
the spec
... approx schedule?
Mark: Unfortunatly not - might help if Steven jumps in as well
Steven: Sean and I have some 5
drafts on the edge
... I'll try to bring it forward
Ben: Is it possible by end of week to have the status?
Steven: Will put it on the HTML-WG agenda
Ben: Good. Please keep us up-to-date
<RalphS> scribenick: ralphs
<michaelhausenblas> http://www.w3.org/2006/07/SWD/wiki/RDFaTestsuite
-> http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2007Feb/0013.html [RDFa] Test Cases [Michael]
Michael: I raise several
questions:
... + What is the host (HTML, XHTML, etc.) for RDFa?
... + What parts of the spec do we test?
... + How do we partition tests? cf. [1]
... + Will we have a (semi-)automated test-environment?
... + Technical issues (test result format: N3, NTriple,
etc.)
... + Legal issues (license)
Steven: what we are testing is whether the page generates the triples we expect, right?
Michael: yes, but we're testing
the transformations
... it makes sense to align with the RDF test cases
Steven: W3C has two sorts of test
suites;
... the majority test the _specification, some test
_implementations_
... we should decide which of these we want to do before we
decide how to go about it
Michael: per the QA test development FAQ terminology, I think we're going for conformance testing
Steven: one of the questions we will need to answer is whether the spec can exit CR
<michaelhausenblas> http://www.w3.org/QA/WG/2005/01/test-faq#why
Steven: we need to show
interoperability test results for that
... to test the specification, for each part we would identify
the assertions and produce a test showing whether an
implementation really does this
... what QA calls 'conformance testing'
Michael: what about the host language?
Ben: our task is RDF in
XHTML
... so I think our tests should be hosted in XHTML 1.1
Steven: sounds good
Mark: agree
<michaelhausenblas> so it reads XHTML 1.1 + RDFa module
Michael: are we chartered to produce a test suite?
Ben: yes, SWD WG is chartered to produce an XHTML 1.1 module
Ralph: test suites are a general expectation, there will not be any objection [based on charter] to us working on it
Ben: add tests [from the RDF test suite] for those features we know we're going to support
<michaelhausenblas> http://www.w3.org/2006/07/SWD/wiki/RDFaTestsuite/TestTypes
Ben: later add others as we resolve open issues of which RDF features to include
Michael: on partitioning of
tests, I took the structure from the RDF Syntax document
... or we could do some other partitioning
Ben: [following Syntax document] looks like a good partitioning
Steven: for CR testing, following
the syntax doc works best
... tests can be explicitly linked to the part of the spec
you're trying to test
Ben: it's harder to test my javascript library against conformance tests than it is to test Elias' library
Michael: Elias' extractor should be tested against our test suite
Ben: the difference is that
Elias' extractor have a REST interface
... other implementations, like my javascript implementation,
don't have REST interfaces
... we could propose automated testing for things that do have
REST interfaces
Mark: what about SPARQL-like
tests?
... "given this input, here's what you should return"
... i.e. define our tests in terms of SPARQL
Michael: is SPARQL close enough to REC?
Ralph: yes :)
Mark: don't require implementing all of SPARQL but could be a way to show that the outputs are the same even if they come from different environments
Ben: I'm not sure this solves my
problem; my implementation runs inside a browswer
... I don't know how to drive a test in a way that would
respect javascript cross-domain limitations
Mark: I'm not suggesting that you
retrieve data in the same manner as SPARQL
... make a test file that contains lots of RDFa and a reference
to your script library
... the script runs and creates a bunch of RDF output, then we
query that output and compare the query results against another
implementation
Ben: there's still some custom
code I have to write for my implementation
... it's more complicated for us to specify how to do this than
for Elias' libraries
... possible, just a lot more work
Michael: propose that for the first phase we only do semi-automatic interface for REST implementations
Mark: I mention SPARQL because
it's not clear what REST has to do with our work
... could say instead that if you run a query over the output
of an RDFa tool you get these triples
... has nothing to do with the transport
Ben: the transport is just a test
implementation method
... the javascript RDFa implementation is meant as a personal
[tool]
Mark: understand that. If you
think of this as a mathematical proof, how do we prove that the
expected triples are generated?
... it's a different thing to say that querying the result of
RDFa extraction produces the following result
... this allows a processor to produce additional triples that
we might not know about
... using SPARQL could hide a lot of implementation details
Ben: sounds like a different
question beyond automation
... you're proposing that we use SPARQL to automate tests
rather than REST
Mark: yes. I'm thinking about
this because I've been playing around with Ivan's examples for
the XML literal issue
... it's easy to use SPARQL to decide if Ivan's examples
conform
Ben: have to explain to
developers how to produce RDFa
... may be too much to have to tell them they need to buy-in to
the whole RDF world
... seems easy to say "it's just a big RDF triple store which
you query" but that may be too much
Steven: just say 'if you query
it, you get this' doesn't leave much to test
... like looking for P element
Mark: would something fail if it produces 12 triples instead of 10?
Steven: no, we could just say that the implementation has to produce _at least_ these triples
Mark: the javascript bookmarklet could define the structure of json but that would be very implementation-specific
Michael: I'd like to prepare a
GUI sketch of an interface to a tester for next week
... could then consider this in the light of SPARQL
<scribe> ACTION: Michael produce a GUI sketch of a test interface for 12-Feb meeting [recorded in http://www.w3.org/2007/02/05-htmltf-minutes.html#action11]
Ben: my main worry is making sure
that RDFa is simple
... simple enough for people to start adopting it
... we don't want to force everyone to read every document
published by the Semantic Web Working Groups
Mark: there can't be half
measures about test suites
... on the other hand, Joe PublicAuthor only needs a nicely
laid out [HTML] page that shows what to expect; he doesn't need
to be involved with a test suite
... microformats current lack a validator; there's no way to
validate that you get what you expect
<michaelhausenblas> http://xmlarmyknife.org/docs/rdf/rules/
Michael: agree ...
<benadida> scribenick: benadida
Michael: NTriples might be easier
to test
... legal issue, which license? Karl D. could help us there
Steven: W3C software license
Michael: we can postpone the
legal discussion
... Wing and I will start working on the details of the test
suite this week
<michaelhausenblas> Scribe: Michael
<Steven> http://www.w3.org/Consortium/Legal/2002/copyright-software-20021231
<michaelhausenblas> Continues via mail
<michaelhausenblas> Michael will look at SPARQL issues
Mark: issue with language
... for equivalences
<michaelhausenblas> next telecon