See also: IRC log, agenda
<DanC> I 2nd the motion (from the agenda) to approve minutes 2 May http://lists.w3.org/Archives/Public/public-grddl-wg/2007May/att-0009/02-grddl-wg-minutes.html
<HarryH> RESOLVED: to approve GRDDL WG-- 02 May 2007 as a true record
<scribe> Scribe: john-l
<DanC> I'm available to meet next week, 23 May... er... I'm getting on a plane in the evening...
<DanC> (I see an agenda request to discuss "base test cases")
<HarryH> to meet again Wed, 23rd May 11:00-0500.
<HarryH> RESOLVED: to meet again Wed, 23rd May 11:00-0400
<DanC> (looking at http://www.w3.org/2004/01/rdxh/spec#changes )
<HarryH> ACTION: for jjc to look at namespace document closely. [recorded in http://www.w3.org/2007/05/16-grddl-wg-minutes.html#action01]
<HarryH> DanC gets victory on actions otherwise.
<DanC> ACTION: chime to update dom in ack [DONE] [recorded in http://www.w3.org/2007/05/16-grddl-wg-minutes.html#action02]
<DanC> (that was done a week or two ago)
<DanC> I think I updated http://www.w3.org/2001/sw/grddl-wg/td/test_results recently...
<DanC> yes... during tutorial prep... 1.13 $ of $Date: 2007/05/06 00:07:08
<DanC> (anybody seen raptor results using the modern EARL namespace etc.?)
<HarryH> Test-Cases that 2 implementations are not interoperable with:
<HarryH> htmlbase3, htmlbase4
<HarryH> xmlbase1, xmlbase2, xmlbase3, xmlbase4
jjc: The subgroup formed to
approve certain tests decided that GRDDL results are
"encapsulated" (according to RFC 3986) within the source
document at the root element of that document.
... I'm not sure if that led to a reapproval or a removal of an *HTML* base test.
jjc: The case in question is #htmlbase1
jjc: The key question is "what is the correct subject of that statement?"
<jjc> <base href="http://www.w3.org/2001/sw/grddl-wg/td/xhtmlWithBaseElement"/>
jjc: (From within the source
... We need to find the base URI of the GRDDL results...
... and for this we turn to RFC 3986.
DanC: The GRDDL spec is *not* "normatively silent" on the base URI for GRDDL results.
jjc: So what does the spec say?
<DanC> "Applying this transformation to po-doc.xml yields RDF/XML; we parse this to an RDF graph (using the URI of the source document, http://www.w3.org/2003/g/po-doc.xml, as the base URI) "
<jjc> The base IRI for interpretting relative IRI references in a serialization of a graph produced by a GRDDL transformation is the IRI of the source document.
<DanC> (it should be in http://www.w3.org/2004/01/rdxh/spec#rule_txprop )
<jjc> "RESOLVED: Given that a base URI parameter is a parameter whose value is the base URI of the source document, the WG RESOLVES not to define a base URI parameter for transforms." (quote of previous resolution)
Seems like this pretty much invalidates the base URI test cases that the subgroup approved.
<HarryH> Groups can always retract approval and re-approve the fixed test-cases.
<DanC> the text jjc quotes ("The bse IRI...") is from 1.233 (connolly 01-Mar-07)
<jjc> 5.1. Establishing a Base URI . . . . . . . . . . . . . . . . 28 5.1.1. Base URI Embedded in Content . . . . . . . . . . 29 5.1.2. Base URI from the Encapsulating Entity . . . . . 29 5.1.3. Base URI from the Retrieval URI . . . . . . . . 30 5.1.4. Default Base URI . . . . . . . . . . . . . . . . 30 5.1.1. Base URI Embedded in Content
GRDDL.py is another implementation that is closely tracking the spec.
<jjc> "Within certain media
types, a base URI for relative references can be embedded within the
content itself so that it can be readily obtained by a parser. This can
be useful for descriptive documents, such as tables of contents, which
may be transmitted to others through protocols other than their usual
retrieval context (e.g., email or USENET news).
It is beyond the scope of this specification to specify how, for each media type, a base URI can be embedded. The appropriate syntax, when available, is described by the data format specification associated with each media type."
<jjc> <base href="http://www.w3.org/2001/sw/grddl-wg/td/xhtmlWithBaseElement"/>
<DanC> (the http://www.w3.org/2004/01/rdxh/spec#rule_txprop rule assumes you have an RDF/XML document, and those have base-URI built-in)
Do we want xml:base and html:base declarations to affect GRDDL results or not?
scribe: that is, declarations within the source document?
jjc: I think that tests currently resolve relative references using base URI information on the root element of the source document.
<HarryH> ACTION: john-l to implement htmlbase1. [recorded in http://www.w3.org/2007/05/16-grddl-wg-minutes.html#action03]
<DanC> I think "The base IRI for interpretting relative IRI references in a serialization of a graph produced by a GRDDL transformation is the IRI of the source document." is broken in that (a) it's not marked as a normative assertion, (b) it should speak of the root element rather than the document.
I agree at least with (b).
<HarryH> ACTION: jjc to write text about base-uri, which may become normative. [recorded in http://www.w3.org/2007/05/16-grddl-wg-minutes.html#action04]
<HarryH> I agree with (a) and (b).
<HarryH> ACTION: john-l to turn "fails" into passes on GRDDL.py and rebuild test-results, bonus points for Raptor. [recorded in http://www.w3.org/2007/05/16-grddl-wg-minutes.html#action05]
<jjc> On IRI text I think (c) it should mention *base* IRI of root element
<DanC> ACTION: chimezie to start an index of tests by normative assertion (proxy for feature) [CONTINUES] [recorded in http://www.w3.org/2007/05/16-grddl-wg-minutes.html#action06]
<jjc> (john note there is some SPARQL code that turns fails into not aplicables where appropriate)
Ah, good to know, thanks. In CVS?
<DanC> (push to agenda review... checking that actions from last time were carried forward http://lists.w3.org/Archives/Public/public-grddl-wg/2007May/att-0009/02-grddl-wg-minutes.html#ActionSummary )
<HarryH> ACTION: [CONTINUES] harry to start an index of tests by issue (not urgent; due 30 May) [recorded in http://www.w3.org/2007/05/16-grddl-wg-minutes.html#action07]
<HarryH> David Booth had minor editorial comment.
There was a fair chunk of discussion surrounding the old resolution of the faithful infoset issue.
In this discussion, several members of the group informally agreed that mapping from XML serialization to an XPath root note is a problem larger than the scope of the GRDDL charter.
Some of this discussion indicated that D. Booth might want to take his comment to the XProc working group.
<HarryH> ACTION: jjc to ask David Booth to reply to current messages on comment list and possibly, if needed, clarify what he wants if he isn't happy. [recorded in http://www.w3.org/2007/05/16-grddl-wg-minutes.html#action08]
<jjc> Jeremy notes that David has been holding off because he has been discussing off-list with jeremy
<HarryH> "The failure of a transformation could be interpreted by a processor as producing no RDF data (fail silently) or the error could be passed back to the invoker. On producing non-RDF elements, let me extend your analogy and say that a processor returns both 11 and "volunteer". The extraneous value could be silently ignored, or the entire answer could be considered meaningless and the invoker informed of the error. We are suggesting that both cases be acknowledged in the specification, and that you specify the behavior allowed by a GRDDL processor."
DanC: He didn't give any reason why we should address his issues.
<jjc> The extraneous value could be silently ignored, or the entire answer could be considered meaningless and the invoker informed of the error.
DanC: What about having a test case addressing the issue?
jjc: We might need additional test machinery.
DanC: These could be informative tests.
jjc: Introduce them with text like "The GRDDL spec doesn't specify what happens in these cases, but..."
<jjc> "some grddl implementations behave as follows"
jjc: Some implementations might get partial results.
DanC: But if only some triples get added to the GRDDL result, that would be a mistake.
HarryH: So a normative test case would solve this problem.
<HarryH> Correct? jjc to write an normative test case showing that if the transform is broken, you should not get a partial result, you should get no results.
<HarryH> ACTION: jjc to write an normative test case showing that if the transform is broken, you should not get a partial result, you should get no results. [recorded in http://www.w3.org/2007/05/16-grddl-wg-minutes.html#action09]
<HarryH> Will talk about 3 remaining comments later.
DanC: Room was packed for the
... 60 people, good timing, audience followed along.
HarryH: People liked the spreadsheet example in Fabien's presentation.
<DanC> (the URIs I've seen from Fabien seem to have /tmp/ in them. anybody have a cool URI?)
HarryH: Companies mentioning the desire to integrate their data, which is often largely stored in spreadsheets.
HarryH: Spreadsheets only came up when raised during the Q&A section.
<HarryH> ACTION: HarryH to look in the spreadsheet into RDF in primer. [recorded in http://www.w3.org/2007/05/16-grddl-wg-minutes.html#action10]
<HarryH> Anyone want to write Excel XML to RDF transform?
<HarryH> Calconnect = www.calconnect.org
<HarryH> Add Excel to Primer
<HarryH> Lee Feigenbaum did Spreadsheets to RDF.
<HarryH> in Dev track.
<DanC> ah yes... http://www.calconnect.org/tcs-active.shtml ... they've been in contact via www-rdf-calendar several times. Matt May attended one of their events in person. I think I blogged about them once.
<jjc> lists about four docs which appear tp be iCalendar
<HarryH> ACTION: Danja to pick LeeF over Spreadsheet to RDF examples, does he have XSLT? [recorded in http://www.w3.org/2007/05/16-grddl-wg-minutes.html#action11]
<HarryH> Meeting Adjourned.