TAG Weekly Teleconference

5 Apr 2005


DanC, Norm, noah, HT, TimBL, VincentQ, Dave_Orchard


See also: IRC log, agenda

Roll call, agenda review

<noah> definite regrets for next week for me.

VQ: [partial regrets?] from EdR

partial regrets from HT

partial regrets from DaveO

timbl: TAG directions... let's talk about that when we have more people

(if anybody thinks it's important, let him make some progress by email, I said. otherwise, let's wait 'till ftf)

<noah> Noah: agree with Tim. Even if we can't fully settle our agenda, early discussion on the phone/email may be useful.

-> minutes 29 Mar

RESOLVED to approve.

Date of Next telcon: propose Tuesday 12 April 2005 -- Scribe, Regrets

Regrets Noah for 12 Apr

regrets timbl, ht for 19 Apr

<noah> Noah is at risk for the 19th

<scribe> ACTION: VQ to find a scribe for 12 Apr [recorded in http://www.w3.org/2005/04/05-tagmem-minutes.html#action01]

<scribe> [PENDING] ACTION Henry: with help from Ed to draft proposal on where in date space to put minutes

TAG Summary for AC meeting

<scribe> ACTION: VQ to draft TAG summary; due end of April [recorded in http://www.w3.org/2005/04/05-tagmem-minutes.html#action02]

VQ: schedule proposed by Ian seems OK? [several]: yes

<Norm> I will not be at the AC meeting

<noah> Noah will probably not be at AC meeting

DaveO: I plan to be at the AC meeting

DanC plans to *not* be at there

<noah> in part because TAG is no longer co-located

TimBL: I'll be there for most of the AC meeting, but not 2nd afternoon

HT: I'll be there at the AC meeting in France

VQ: I'll be there too

DaveO accepts HT's nomination to present to the AC, working with VQ

namespaceDocument-8, RDFinXHTML-35

. ACTION DanC own (i.e. propose to close) issue RDFinXHTML-35

DanC: some progress... please continue

<scribe> [PENDING] ACTION DanC: own (i.e. propose to close) issue RDFinXHTML-35

. ACTION Norm to pick up Paul Cotton's work on namespaceDocument-8

-> http://lists.w3.org/Archives/Public/www-tag/2005Apr/0003.html ndw on namespaceDocument-8

NDW: I sent that "is RDDL 1.0 it?" msg; I'd like to get more feedback

TBL: how does GRDDL interact? does standardizing it moot standardizing RDDL?

NDW: hmm... yes, we could say "use any concrete syntax you like, as long as you map it to this model"

NM: [something about popular somethings... darn. missed it]

TBL: [...] yes

NM: then I'd put the emphasis on the ontology, since that's what people would have to agree to
... and we should give some guidance as to how much variability to encourage

DC: re "presumably somebody has GRDDL'd RDDL" ... well, that's in progress... and it brings up issues of sharing fragments across HTML/RDF

HT: RDDL is deployed in several schema validators; they'll look thru RDDL documents for schemas

<Zakim> timbl, you wanted to refine on Noah's comments about h"the ontology" to talk about "an ontology".

HT: this "GRDDL obviates RDDL" idea involves adding XSLT processing and RDF/XML parsing to that mix

TBL: yes, the "you need XSLT to grok any namespace document" point is an interesting one... but as to noah's point about ontology...
... while the TAG might have some expectations about what data a namespace document should provide... say persistence... the point of communicating in RDF is that you can use any ontology, e.g. community-specific metadata

HT: yes, RDDL likewise had an extensible palette [sp?] from which to choose... DTDs, schemas, stylesheets, etc.
... it seems important to distinguish documenting namespaces from documenting languages
... namespaces and languages aren't 1-1
... e.g. html, which is one namespace, 3 languages
... and things that axiomatize the rendering model interact with each language (though there are some commonalities)
... and consider 2 different versions of XSLT, which have the same namespace but different datamodels

<Zakim> noah, you wanted to say: yes, I knew the ontologies are inherently extensible, but...

NM: yes, ontologies are extensible, but like the RDDL "standard library" of choices, we should choose some

<scribe> ACTION: NDW to take GRDDL/RDDL discussion to www-tag to solicit feedback on directions for namespaceDocument-8 [recorded in http://www.w3.org/2005/04/05-tagmem-minutes.html#action03]

HT: I've been trying to write up this namespace/language stuff... I submitted an abstract to a conference... so if it's accepted I'll be writing some more

DaveO: sounds interesting

(yes, quite relevant to xmlVersioning-4x)

Educational materials

VQ: continuing discussion from Boston meeting...

. ACTION DanC to make an "educational materials" page, describe what we're interested in, and set expectations

-> http://www.w3.org/2001/tag/em27.html education materials page

<noah> I think this looks great as a start.

NM: how about a link to webarch?

DC: oh yeah.

TBL: with a summary... say that webarch has been thru lots of review, has examples, fairly technical, about 30 pages

some discussion of level of announcement... maybe just www-tag for now, louder later

<scribe> ACTION: DanC to add a webarch link, spell check and the like, and announce "educational materials are welcome" to www-tag [recorded in http://www.w3.org/2005/04/05-tagmem-minutes.html#action04]

issue URNsAndRegistries-50

. ACTION Ed to review XRI and provide comments

VQ: I got an extension thru end of April

. ACTION Henry and David to draft initial finding on URNsAndRegistries-50

HT: I have some comments from DO that I haven't integrated
... I noticed that the XRI group moved back their review even though it cost them a month[?]

DC: HT, found a thesis statement?

HT: "HTTP URIs are good names; they meet all the requirements"

TBL notes the abstract says basically that

(no, henry, you can't ask to postpone the technical discussion by making a technical point!!!)

HT: people say "we need something other than http: because http: leads to 404"; the answer seems to be "anything else will have the equivalent of 404 too"... I wonder if there's a better answer

DanC: perhaps more elaborate, but not much better, no

<scribe> [PENDING] ACTION Henry and David to draft initial finding on URNsAndRegistries-50

<scribe> [PENDING] ACTION Ed to review XRI and provide comments

binaryXML-30, XML Binary Characterization

VQ: XBC WG deliverables have been delivered; WG is now closed, but we're still on the hook to contribute to community discussion (esp team, AC) continues

<Norm> : My review of the use cases document for XML Core

. HT to review the XBC documents with help from Noah
. Ed and Norm to review XML Binary Characterization use cases document.

VQ: I gather the team wants to get something together for the AC in June

DC: yes, that's what I hear

yes, http://lists.w3.org/Archives/Public/public-xml-binary/ is public

VQ: we can comment individually... shall we comment as a group?

DC: yes, please, let's aim to resolve binaryXML-30 before June

NM: well, i was going to work with HT in the 2 weeks after Boston... the next 3 weeks are pretty much "worst case" w.r.t. my availability

(NM and DO discuss review of some draft)

(ball is in NM's court)

<Zakim> Norm, you wanted to say that the use cases didn't leave me feeling very informed and the xbc characterization document has some significant editorial problems

<Norm> My favorite comment from the characterizatoin document is that XML "prevents processing efficiency" and "prevents space efficiency" and "prevents forward compatibility"

<Norm> I concede that "unconvinced" would have been a more apt characterization than "uninformed".

at this point, DanC got conflicted between scribing and participating. TimBL voiced most of the comments below.

DanC: my impression of at least the TV metadata use case is that it argued as strongly for text-XML as binary-XML

<timbl> I felt the arguments were listed without being analyzed and judged.

<timbl> So one is not left with a recommendation.

<timbl> I'm not sure that they aimed to do that.

<timbl> What I would have liked to have seen more analysis, such as a breakdown of classes of solution against these.

<timbl> http://www.w3.org/TR/2005/NOTE-xbc-use-cases-20050331/

<timbl> I was concerned that each use case was not compared with compression.

NM: so there's more text than when the binary XML workshop was held, but I'm frustrated that these drafts don't make the case... maybe don't even try

DO: BEA's position at the binary XML workshop was "we'd prefer 0 new formats to 2 new formats" and I'm still not convinced that any one format is going to help in an interesting set of use cases

<timbl> I was also concerned that there were no performance figuresd

<noah> To amplify on Dan's scribing of what I said: my understanding is that the characterization workgroup has concluded that not only are the many use cases important, but that there is a single technology (not identified) that could meet them all. I don't think they've justified that conclusion.

TBL: all this talk of performance with no numbers is very counter to my engineering experience
... there were some CPU megahertz figures, but nothing that said "and so it's X% slower in parsing XML in text than binary"

<noah> I completely agree with Tim. Not only are careful measurements essential, it's important to do comparisons with the best availalble parsing technology for text XML. Many parsers such as Xerces were written for correctness, flexibility, and quick time to market as opposed to blazing performance.

<timbl> I wonder in passing whether XML parsers for a simpler XML would be faster.

Reviewing the WS-Addressing Core and SOAP Binding documents

VQ: we got mail about the WS-addressing last call schedule, soliciting review

<timbl> http://dev.w3.org/cvsweb/~checkout~/2004/ws/addressing/ws-addr-core.html

( http://www.w3.org/2001/tag/issues.html?type=1#endPointRefs-47 )

TimBL: I'm in contact with PLH about some standards process issuess... WSDL/SOAP versions...

<timbl> normative recommendations to non-standards

DC: exploring the null hypothesis... what if we close endPointRefs-47 with no action, just "never mind"?

(long quiet)

TimBL: it's consistent with saying "we don't see a coherent architecture on which to comment"

DC: I don't understand ws-addressing... it seems that every document is legal; I don't see how you could fail a test

VQ: losing energy... ready to Adjourn?

DanC, scribe, for VQ and the TAG
$Revision: 1.4 $ of $Date: 2005/04/12 17:09:30 $
Minutes formatted by David Booth's scribe.perl version 1.122