GRDDL -- 18 Jul 2007

Convene GRDDL WG meeting of 2007-07-18T11:00-0400

<HarryH> Who wants to scribe?

<HarryH> DanC is running late.

i can scribe

<HarryH> Thanks David.

<scribe> Scribe: DBooth

<rreck> i mentioned how great GRDDL is at my paper last Saturday

<danja> rreck, cool

good!

<HarryH> PROPOSED: to approve GRDDL WG -- 11th Jul 2007 as a tru e record

<HarryH> http://lists.w3.org/Archives/Public/public-grddl-wg/2007Jul/att-0017/11-grddl-wg-minutes.html

<rreck> i told encyclopedia britannica about GRDDL today

<HarryH> RESOLVED: approved GRDDL WG -http://lists.w3.org/Archives/Public/public-grddl-wg/2007Jul/att-0017/11-grddl-wg-minutes.html - 11th Jul 2007 as a true record

<HarryH> PROPOSED: to meet again Wed, 25th July 11:00-0400. scribe volunteer?

<rreck> regrets for next week, i will be in lancaster

<HarryH> In Toronto but can probably make it.

RESOLUTION: to approve GRDDL WG -- 11th Jul 2007 as a tru e record http://lists.w3.org/Archives/Public/public-grddl-wg/2007Jul/att-0017/11-grddl-wg-minutes.html

<Simone> regrets also for me for me, i'll be in prague

RESOLUTION: to meet again Wed, 25th July 11:00-0400, Scribe DanC?

Standard Library

<HarryH> Standard Library Transform

John: Found description of attributes confusing, but maybe only applies to trick cases like Jeremy's example. The issue of dealing with URI base is scary.

Jeremy: 3 ways to go, not clear which is best.
... I need to respond to John. ONe choice is the attribute is not needed. It's tested in the lib test but not in the others.

DanC: Why scary?

John: You can get yourself into cases where you need to throw a terminating error.

Jeremy: We had a test suite case for embedded RDF. Not sure of status -- I think not approved. Had two relative XML bases above it, and it was too difficult to get right without proper XSLT processing that I was trying to avoid.

John: I think a GRDDL transform that gives up in the middle is problematic.

DanC: Not that we cannot produce GRDDL results, but we chose not to. What's the problem?
... The purpose of the lib is to help people do straight-forward things -- not crazy things.

Jeremy: Code that implements full URI resolution would not be maintenance-free code, and I don't think the GRDDL lib code is something we should expect money should be spent to maintain.

Harry: If it does help GRDDL transform authors it's better than nothing. There's some sort of error message it produces. As long as that is very clear about what is going on.

Jeremy: It's reasonably clear.

DBooth: Should the error say more about why we're not implementing more?

Jeremy: If so it should point to more documentation.

<HarryH> TION: jjc to organize library tests; link from http://www.w3.org/2001/sw/grddl-wg/library [CONTINUES]

<HarryH> ACTION: jjc to organize library tests; link from http://www.w3.org/2001/sw/grddl-wg/library [CONTINUES] [recorded in http://www.w3.org/2007/07/18-grddl-wg-minutes.html#action01]

<HarryH> ACTION: jjc to dispose of @@s in http://www.w3.org/2001/sw/grddl-wg/library (and in docs of 3 txforms) [CONTINUES] [recorded in http://www.w3.org/2007/07/18-grddl-wg-minutes.html#action02]

Jeremy: I'm happy to add a pointer to more explanation from the error message.

<inserted> John: The error message seems clear enough.

GRDDL Outreach

Harry: I saw people playing with Earl results from openlink.

<HarryH> ACTION: Danja to contact Kingsley and try to get GRDDL EARL results [DONE] [recorded in http://www.w3.org/2007/07/18-grddl-wg-minutes.html#action03]

DanC: I put them in the test results document. They were not bad.

<jjc> <xsl:text>http://www.w3.org/2003/g/xml-attributes

<jjc> Nested relative xml:base attributes are not supported.

<jjc> The relative base URI "</xsl:text>

<jjc> <xsl:text>" presented problems, maybe replace with an absolute URI.</xsl:text>

DanC: They didn't report on HL7 to OWL. Passed all informative tests. Said something about how their XSLT implmentation is incomplete, and defers to the W3C implementation when it craps out. I think that's a temp limitation.

<DanC> http://www.w3.org/2001/sw/grddl-wg/td/sponger.rdf

<HarryH> ACTION: IanD and Danja to e-mail maintainer of Dublin Core Metadata Profile to upgrade to GRDDL [CONTINUES] [recorded in http://www.w3.org/2007/07/18-grddl-wg-minutes.html#action04]

Harry: Danny, apparently Andy Powell coming to your offices tomorrow?

Danny: That's the first I've heard!

Harry: Best RDF transform for ATOM?

Danny: Havent' looked closely yet. The output of David Powell's earlier version was painfully verbose.
... Kanzaki one is very nice.
... Three candidates. Comes down to personal pref. 2 for XSLT1 and 1 for XSLT2.

Harry: Just need to choose one transform. I would guess XSLT1 would be the best choice.

DanC: I want one that's done.

Danny: The Kanzaki one looked pretty good but I didn't look very closely.

<DanC> (to clarify, I don't think the IETF AtomPub WG is likely to be interested in somethingt that's not done.)

Harry: Could prob use another pair of eyeballs on this.

Danny: I could probably run Kanzaki's test cases against the others.

DanC: What namespace is RDF output?

Danny: I think they use Atom namespace with a hash at the end.

DanC: I know at one point Henry Story was using a made-up W3C URI, and I told them "get off our lawn".

Harry: Anyone else want to noodle this?

John: I'm interested. But how to get this by the Atom WG?

Harry: Hello WG, this will help Atom interact with RDF. Like it?

John: Asking them to bless it, or add a reference to a namespace?

Harry: A ref to namespace. But the point is that using it is optional.

Danny: Would have to carefully worded.

DBooth: Would be helpful to privately probe to find out if there's gong to be resistance.

DanC: I'm willing to ask Tim Bray once there's something more solid.

John: Should we allocate some GRDDL space for this on CVS?

Harry: Makes sense.
... In the TD directory?

Danny: Not sure of licensing.

DanC: Different dir would be better.

<danja> can check easy enough

<DanC> atg18

<DanC> for atom grddl, 19 july

<DanC> going once?

<DanC> er... 18 july

<DanC> $ cvs add atg

<DanC> Directory /w3ccvs/WWW/2001/sw/grddl-wg/atg added to the repository

DBooth: atg?

DanC: Done.

<HarryH> ACTION: danja to find out the status of Atom GRDDL and IETF Atompub WG list to find other people with transformations or schemas.[DONE] [recorded in http://www.w3.org/2007/07/18-grddl-wg-minutes.html#action05]

DBooth: Too early to probe atom folks?

DanC: Tim Bray yes.

Danny: Others too.

DanC: What about the transformation to RSS1? -- hugely political

DBooth: I suggest we not touch that at least until we've probed about conversion to RDF.

DanC: In some ways this would be throwing in the towel on RSS1.

<danja> before making any choice I'd like to canvas the authors of each of the current candidate XSLTs (& dajobe) for their opinions

DanC: I'd prefer one std for syndication. Seems easy to migrate to a new namespace.

DBooth: i agree with Danny's IRC comment. I'd rather not risk raising their ire.

<danja> yep, Nature & iand too - they know their stuff with RSS 1.0

<HarryH> ACTION: danja and john-l to test and see if any XSLT is mature enough to add to ATOM namespace doc and has proper licensing (W3C Software license), as well as canvass all candidate XSLT authors for their opinion on ATOM/RSS 1.0 namespace.. [recorded in http://www.w3.org/2007/07/18-grddl-wg-minutes.html#action06]

Harry: Curious about who is interested in converting microformat data into RDF.
... There are 2-3 blank spots.

DanC: What would the survey be for hcalendar?

<rreck> working on it as in modifying the xslt?

Harry: Who would be interested? Head count.
... I know Tom Grover's interested. But I'd like a head count. If we can't get hcal, then maybe profile URIs.

<DanC> sure, scaring up people to work on hCalendar/RDF testing would be great.

<HarryH> ACTION: HarryH to make Questionnaire as regards microformat XSLTs and e-mail swig list. [recorded in http://www.w3.org/2007/07/18-grddl-wg-minutes.html#action07]

Relationship to XHTML 2 and HTML 5

<DanC> (taking a peek at kanzaki's work, I see in the output: <rss:channel rdf:about="http://alexharden.org/blog/atom.xml"> )

<HarryH> http://lists.w3.org/Archives/Public/www-tag/2007Jul/0048.html

<HarryH> http://www.w3.org/2007/05/14-html-wg-irc#T19-47-16]

<HarryH> http://www.w3.org/2007/05/14-html-wg-irc#T19-47-16

Harry: There's been a lot of talk of profile dropping.
... MarkBaker's saying we should use something other than profile. I'm not sure why.

DanC: Neither of you seems to be saying anything about URI extensibility.

Harry: Nobody in the TAG discussion is saying that's bad. Only in the HTML5 discussion.

DanC: I think Mark's being kinda helpful. He's saying "that argument is not compelling" -- what the other side may say.

Harry: I'm not terribly interested in extending profile, only in keeping GRDDL working. Is mark serious about extenting profile?

DanC: I think he's playing devils advocate.

Harry: What else happend in that discussion?

<danja> (phew, the HTML5 irc log is heavy going...)

DanC: I'm trying to set expectations, but the latency is months. HTML group will get to it eventually.

Harry: Since GRDDL wg won't be here, how should we track this? DanC, could you pop an email to TAG when the issue comes up in HTML wg?

DanC: Yes.

<jjc> on extensibility

<jjc> http://lists.w3.org/Archives/Public/www-tag/2007Jul/0051

<jjc> mark says:

<jjc> I don't think centralization is prima facie bad.

<DanC> (if you want an action tracked, one of us could bring this to the semweb-cg)

Harry: Are URIs as the point of extensibility flawed?

DanC: It appeals to me intuitively, but there's been no discussion of it.

Harry: The greater issue seems to be the way HTML5 is writing their spec: the only reason to include something is evidence.

DanC: Not necessarily.

<DanC> (I didn't so much mean "there's been no discussion of it", but rather: I haven't managed to articulate the argument, and I haven't seen anybody else do it either.)

Harry: Cannot appeal to evidence to argue for profile.

DanC: If you can articulate the argument for URIs as extensibility points that would be helpful.

<HarryH> ACTION: HarryH e-mail about @profile to XHTML 2 WG. [recorded in http://www.w3.org/2007/07/18-grddl-wg-minutes.html#action08]

PR approved

<rreck> please repeat that harry

DanC: Do people know how to do testimonials? Do you know who are your AC reps?

<rreck> can i do a testimonial?

<DanC> http://www.w3.org/2002/09/wbs/33280/GRDDL-200707/

(DanC does a poll of who people's AC reps are)

DBooth: I believe HP will be doing one.

DanC: You have your AC rep follow the form above.

<rreck> does everyone else hear wind?

Harry: What high-profile names/companies could endorse GRDDL?

<Simone> Simone: My AC in chat said OK for testimonials to GRDDL and we may implement it on our site

<rreck> i can ask around

<Simone> "IWA AC Rep said: ok, we can make a testimonial comment"

<danja> (iand has mailed in a testimonial, I've just pointed him to the form)

Harry: Could Dave Beckett do a testimonial as an employee of Yahoo?

<HarryH> ACTION: HarryH to ask David Beckett to ask AC rep to a testimonial. [recorded in http://www.w3.org/2007/07/18-grddl-wg-minutes.html#action09]

<danja> "iand: okey dokey"

DBooth: Any google contacts to tap?

Harry: How about Tantek?

<HarryH> ACTION: HarryH to ask RyanKing for Testimonial about microformats and GRDDL. [recorded in http://www.w3.org/2007/07/18-grddl-wg-minutes.html#action10]

DanC: No longer works for Technorati. His blog says he's rededicating himself to microformats.

Ron: I could mention to Encycl. Brit.

DanC: Any gov contacts allowed to say anything in public?

Ron: Lemme think about it.
... Deadline is Aug 24.

<HarryH> ACTION: rreck to ask Ken and Encyl. Brit. for testimonials. [recorded in http://www.w3.org/2007/07/18-grddl-wg-minutes.html#action11]

Ron: I showed Enc Brit people how to take their parser and move it to a more expressive result set.

W3C GRDDL Service

Harry: Can Dom have time to fix this?

<danja> http://people.w3.org/~dom/archives/2007/07/grddl-is-a-moving-along/

DanC: No new data, but he's now Mobile Web activity lead, so may have less time now.
... But he did write a blog entry about GRDDL.

Harry: Also, about schema, there is email that seems to want to acknowledge that the schema need not be at the namespace URI.

DanC: They're saying this is broken. He used the word annotation, and says it doesn't follow third-party annotations.

Harry: I'm concerned that he may not understand that we do have a non-schema way to do GRDDL transformations.

DanC: I don't think so.

<DanC> http://lists.w3.org/Archives/Public/public-grddl-comments/2007JulSep/0002.html

<DanC> Re: Comments on GRDDL (using 3rd-party XML schemas with GRDDL)

Harry: Keep us posted about that.

<danja> (I think DanC's mail should flush out what's needed - I don't personally see any spec changes necessary, but what'll satisfy him is another matter...)

<HarryH> Meeting adjourned.

<HarryH> DanC cannot scribe, DavidB can next week.

<DanC> wow... http://www.kanzaki.com/works/2005/misc/0726atom.html is really good stuff.

s|<HarryH>||

- DRAFT -

GRDDL

18 Jul 2007

Attendees

Contents