See also: IRC log
<scribe> Scribe: DanC
Fabien is excused after xx:20
RESOLUTION: to approve minutes 13 Sep
HH: I may need to leave before end of call
PROPOSED: to meet again 27 Sep, per regular schedule
RESOLUTION: to meet again 27 Sep, per regular schedule, FabienGandon to scribe.
HH: w3c communications team and semweb CG is inclined to do all 3 drafts at once, 24 Oct...
DC: I'm inclined to publish the use cases doc this month, to keep to the 3 month rule
CHI: yeah, let's not sit on docs that are ready to go
HH: ok, I'll run that by the CG: 24 Oct press release date, and publish docs as they're ready
HH: ready to go?
FG: modulo the 2 concerns from
last week...
... but I'm OK to publish with those unresolved.
<harryh> Danny can you give the use-case document a final read? And then give us a response on whether it publishable or not?
<DanC> or have you already read it closely?
<danja> will do - tonight
<danja> (straight after telecon)
FG: on XML Schema, I replied to Dan; I just need one high-level paragraph
CHI: I sent a suggestion.
HH: ok, so it looks like we'll be ready to decide next week
<scribe> ACTION: Murray to send out his take on take on standardized vocabulary [CONTINUES] [recorded in http://www.w3.org/2006/09/20-grddl-wg-minutes.html#action01]
<scribe> ACTION: DanC: with BenA advise FG on CHI's XML Schema paragraph suggestion [recorded in http://www.w3.org/2006/09/20-grddl-wg-minutes.html#action02]
chime's suggestion http://lists.w3.org/Archives/Public/public-grddl-wg/2006Sep/0138.html
<scribe> ACTION: BenA to flesh out HTML RDF/A output of GRDDL use cases [DONE] [recorded in http://www.w3.org/2006/09/20-grddl-wg-minutes.html#action04]
<scribe> see also issue outputformats
<scribe> ACTION: Ryager to read draft thoroughly [CONTINUES] [recorded in http://www.w3.org/2006/09/20-grddl-wg-minutes.html#action05]
<scribe> ACTION: BenA to draft XML Schema use-case in context of Creative Commons and OAI. [DONE] [recorded in http://www.w3.org/2006/09/20-grddl-wg-minutes.html#action06]
ID: I added SPARQL to the primer
ID: I noted DanC's dtend fencepost bug [and fixed it?]
<scribe> ACTION: iand to remove references to RDFa from primer [DONE] [recorded in http://www.w3.org/2006/09/20-grddl-wg-minutes.html#action07]
<scribe> ACTION: iand, Fabien to change wording of POX to 'single-purpose XML vocabulary' or something similar [DONE] [recorded in http://www.w3.org/2006/09/20-grddl-wg-minutes.html#action08]
BenA: what exactly was removed?
ID: a surpious ref
<scribe> ACTION: BenA will proof-read primer document [CONTINUES] [recorded in http://www.w3.org/2006/09/20-grddl-wg-minutes.html#action09]
<scribe> ACTION: chime will review primer document, advise on publishing as WD [CONTINUES] [recorded in http://www.w3.org/2006/09/20-grddl-wg-minutes.html#action10]
<scribe> ACTION: DanC to get IanD cvs write access for primer work in GRDDL WG [recorded in http://www.w3.org/2006/09/20-grddl-wg-minutes.html#action11]
DC: Danny was going to test the primer files... news?
<scribe> ACTION: DannyA to test the primer files [CONTINUES] [recorded in http://www.w3.org/2006/09/20-grddl-wg-minutes.html#action12]
<danja> not done yet, got a bit too involved in the testing code...
DC: hmm... I presented GRDDL last week... sorta...
CHI: the parts of the primer I worked with were OK when I looked
<scribe> ACTION:DanC to enumerate options for base issue, pick one, and propose it in the form of a test case. See his suggestion to pass it as a parameter and discussion with Chime over extension functions. [CONTINUES] [recorded in http://www.w3.org/2006/09/20-grddl-wg-minutes.html#action13]
-> Four options for discussion
HH: we've heard from DanC and Ben; I sent some options...
<danja> I'd go for a variant of 2), SHOULD for RDF/XML, MAY for others (not necessarily W3C-recommended)
DC: I don't have much rationale for requiring RDF/XML other than I have experience and I know how it works.
MM: GRDDL is easy to understand
as "N in, 1 out"...
... I can see it setting a precedent for similar things.
[CHI and DC discussed abstract syntax expressed as API calls rather than any concrete syntax too...]
MM: looking at the use cases and
primer, they remind me of "application notes" about
technologies many years ago...
... I'm concerned that the use cases show doing too much with
GRDDL.
<Zakim> benadida, you wanted to comment on "other implementations"
BenA: if transformations can only output RDF/XML, they can't express the presentation XHTML stuff while they're at it.
<Zakim> chimezie, you wanted to "not really 1 out"
CHI: I prefer thinking of the output as RDF abstract syntax
MM: did I hear that RDF/XML is less expressive than other syntaxes?
CHI: yes
MM: and is there a "maximally expressive" syntax?
<Zakim> iand, you wanted to "too much variability hurts interop"
CHI: well, sort of; SPARQL picked a middle ground rather than, e.g. N3
<danja> without at least one normative concrete syntax, where's the interop?
<Zakim> DanC, you wanted to note that the scope of GRDDL is getting data (think: spreadsheet, database) out of documents.
DC: I lean the same way Ian does; the benefit of allowing other than RDF/XML isn't worth it.
<chimezie> seems like there is a 3 level stratification of GRDDL output 1) a specific RDF concrete syntax 2) any concretized abstract syntax (wouldn't include API calls) 3) any concretized abstract syntax in addition to API calls
<DanC> I see no merit in (2)
<chimezie> I can't even imagine an API call scenario/usecase in the framework we have so far for GRDDL
<iand> problem is that there's no way to predict what syntax is going to be output by any given transformation
<benadida> I think I overstated the need for an API approach in the GRDDL *spec*
HH polls for RDF/XML only vs RDF/XML preferred
HH: sounds about even
<iand> abstract syntax - http://www.w3.org/TR/2004/REC-rdf-concepts-20040210/#section-Graph-syntax
DanC: the even poll makes re-inforces my preference to keep this open until we can get Brian McBride's input.
<iand> this is relevant - http://www.w3.org/TR/rdf-syntax-grammar/#section-Serialising
MM: is there a spec for going from RDF abstract syntax to RDF/XML?
DC: [confuses the issue with his ansewr; iand's pointer is the right answer]
<chimezie> The XML realm is syntactic (uniformity is important), the RDF realm is semantic/abstract (syntactic uniformity is irrelevant)
DC clarifies that RDFa is not the RDF abstract syntax, but rather another concrete syntax
HH: so if we mandated RDF/XML only, and other syntaxes mature, we'd have to rev GRDDL, right?
MM: or make a new thing, maybe "GARDEN". but there's real value in knowing the target of GRDDL.
<chimezie> Somewhat related: The rdf scutters I write (when met with a URL w/out an appropriate mime-type) attempt an RDF/XML parse first, TriX second, N3 third in that order
(lots of discussion, not carefully scribed.)
<chimezie> If there is truely a very minimal set of abstract RDF that can't be expressed in RDF/XML im inclined to rethink my stance on abstract syntax
<iand> http://www.w3.org/TR/rdf-syntax-grammar/#section-Serialising gives you the exceptions
DC: one concrete use case where a transformation designer prefers other than RDF/XML is Atom/Owl. Henry story has written XQuery that produces turtle. I have tried to get him to produce RDF/XML, but I haven't managed.
MM: one possibility is that GRDDL specifies RDF/XML only, but a WG note specifies the less constrained version, that targets RDF abstract syntax
ADJOURN.
<scribe> ACTION:Harry to draft some elaboration on how media types work with GRDDL ("adding a sentence" as in this message). [CONTINUES] [recorded in http://www.w3.org/2006/09/20-grddl-wg-minutes.html#action14]
harryh has made progress, but we haven't discussed the outcome
<scribe> ACTION: Murray to suggest what GRDDL spec issues are covered by XML Processing, suggestion on how to fix it. [CONTINUES] [recorded in http://www.w3.org/2006/09/20-grddl-wg-minutes.html#action15]