Technical Architecture Group Teleconference

12 Feb 2009


See also: IRC log


Jonathan_Rees, Raman, noah, Danc, masinter, John_Kemp, Ashok_Malhotra, DaveO, Stuart
TimBL, Henry



trackbot, start meeting

<trackbot> Date: 12 February 2009

RESOLUTION: to meet again 19 Feb, ashok to scribe

NM: re agenda review, error handlingn didn't work out for this week

RESOLUTION: to approved http://www.w3.org/2001/tag/2009/01/22-minutes

RESOLUTION: to approve http://www.w3.org/2001/tag/2009/02/05-minutes

F2F Content and Agenda Planning (Brief)

NM: our ftf meeting is in 3 weeks...
... 1. progress on substantial topics 2. step back and look ahead for the year
... those are my 2 points/goals re the ftf
... let's have reading materials available well in advance

<Zakim> raman, you wanted to discuss agenda f2f

<masinter> My contribution about TAG priorities

TVR: last Sep, we talked about HTML and tag soup; I'd like Larry and John to look at the way the agenda was put together and what we discussed...
... also, I'd like to know from Tim and Dan what follow-up there has been since the TPAC

<DanC_> (I did report at our Dec FTF on follow-up from the TPAC; I'm happy to do so again)

<masinter> Re HTML5 and "tag soup" etc.: I've tried to read the background material and previous activities here, and frankly, can't really see a "TAG position" here

<scribe> ACTION: DanC report at March on tagSoup progress since TPAC [recorded in http://www.w3.org/2009/02/12-tagmem-minutes.html#action01]

<trackbot> Created ACTION-226 - Report at March on tagSoup progress since TPAC [on Dan Connolly - due 2009-02-19].

NM: we're drifting into technical discussion, which is natural... error handling looks similar...

LMM: error handling is perhaps a bit more broad than tag soup, but yes, it's related

<masinter> it's a generalization of one (but only one) of the issues with HTML5

<Zakim> raman, you wanted to summarize Sep tagsoup for Larry and John

<DanC_> Sep KC meeting record

TVR: the Sep ftf was just before TPAC... so I'd like to see what came out

NM: so it looks like yes, tagSoup should be on the ftf agenda; one session or more?

TVR: let's do something similar to what we did in Sep, with substantial prep

LMM: the TAG has met with the HTML WG... [missed some?] we should have a TAG finding on [did he say all the web standards?]
... for example, should W3C propose accessibility techniques that have not been implemented?

NM: is the ftf a good time to talk about this?

LMM: yes

DO: on XML, HTML... I'm interested in a consensus position of the TAG on HTML, XML, and such; we've discussed it without resolution

DanC: I'm still at-risk to attend the ftf

F2F Logistics (Brief)

NM: Ashok, any logistic details?

<DanC_> March meeting logistics

Ashok: I'll distribute a cell phone to TAG members


<noah> http://www.w3.org/2001/tag/2009/02/05-minutes#item10

close action-224

<trackbot> ACTION-224 Schedule discussion of metadata (scope, issue, coordination, workshop, etc) as item on next telecon closed

LMM: I'm happy to defer this to the ftf agenda

NM: as telcon time is cheaper, maybe better here?

AM: I note a new draft of the site metadata work [by mnot, yes?]

LMM: I think what's needed is a background survey on metadata; I'm hesitant to offer something soon...

AM: Jonathan's piece on metadata use cases is perhaps a good summary

<jar> http://www.w3.org/2001/tag/doc/uniform-access-20090205.html#cross_site

<jar> that has use cases

<DanC_> (gee; evidently I'm behind on email)

<Zakim> DanC, you wanted to wonder about POWDER and site metadata

DC: POWDER let you say things like, everything in this subtree OK for children. Mark Nottingham's stuff as similar capability. Are we the only ones looking at both?

<Zakim> jar, you wanted to ask LM what he had in mind

JAR: I could do something, but I wonder if it's too narrow; I can summarize work by Eran, mnot, [??], ...

<Stuart> fyi... a good summary from Eran is at: http://groups.google.com/group/metadata-discovery/browse_thread/thread/b4f60d20896ad7c5

<masinter> that sounds like a good start

JAR: I wonder if LMM had other stuff in mind; of course metadata is a huge area...

LMM: I've been working on metadata for a couple years, esp Adobe XMP, ...
... I'm interested to reconcile what I've seen with [other stuff?]

<Zakim> DanC, you wanted to suggest LMM look at tag/doc/uniform-access

DC: I think uniform access to metadata by Jonathan and Phil Archer is as good a starting point for overall review as we had.

(request to discuss this cites http://lists.w3.org/Archives/Public/www-tag/2008Nov/0008.html TAG Resolution endorsing W3C Team Comment on the identifications of WS Transfer resources. )

DC: Request to discuss today somehow related to W3C Team work on WS TRansfer. Jonathan do you remember?

JAR: No.

DC: Ashok?

AM: There's a thin connection.
... in WS-Transfer context, they talk about XML representation of resource; to them, a resource is a web service...
... that's one style of metadata...

<Zakim> noah, you wanted to talk about coordinating this

<Stuart> Ashok, are you trying to say (roughly) that the WSRA resourse has (roughly) a schema that might be thought of as resource metadata?

NM: I'm interested to have 1 or 2 tag members to be point-person on issues/areas...
... jonathan is offering to do something... I wonder about expanding it to an overall tracking role...

<Ashok> Stuart, The metadata they talk about is usually XML Schema, WSDL etc

JAR: I'm willing to try, though I wonder about mixing my biases in with TAG priorities...

<masinter> I can work with JAR offline

good, masinter

<Stuart> Ashok... thx.

. ACTION: jonathan summarize TAG work on metadata, with Larry

<Ashok> Stuart, Typically Schema, WSDL and Policy (forgot abt policy last time)

<scribe> ACTION: jonathan summarize TAG work on metadata, with Larry [recorded in http://www.w3.org/2009/02/12-tagmem-minutes.html#action02]

<trackbot> Created ACTION-227 - Summarize TAG work on metadata, with Larry [on Jonathan Rees - due 2009-02-19].

johnk: we should try to become clear on what we mean by "metadata"; the definition of metadata isn't clear enough

LMM: my sense is that this merits W3C work, not just a TAG issue, so I hope the TAG will prompt W3C work

NM: Dan, care to comment as staff comment?

DanC: it's straightforward for the TAG to suggest new W3C work, though starting things in W3C is naturally a very involved process

NM: I have some hesitation... I think the TAG has good standing to suggest others do work >after< we've done some serious work understanding why there's a problem somewhere. Not clear we've done that here. But I'm happy to see what Jonathan and Larry come up with

LMM: I acknowledge that just saying "this looks interesting" isn't useful, but it seems that there are architectural issues that none of the individual efforts are addressing, and it's the TAG's role to coordinate in those cases


<trackbot> ACTION-227 -- Jonathan Rees to summarize TAG work on metadata, with Larry -- due 2009-02-19 -- OPEN

<trackbot> http://www.w3.org/2001/tag/group/track/actions/227

<Zakim> jar, you wanted to discuss reading list deadline for f2f

<DanC_> +1 2 weeks before

JAR: we had a 2 weeks-before goal in previous ftfs...

NM: I mentioned 1 week in advance... in general it depends on size etc.; in this case, 1 week in avance of the ftf seems OK

action-227 due 24 Feb

<trackbot> ACTION-227 Summarize TAG work on metadata, with Larry due date now 24 Feb

Web Services Resource Access (WS-RA)

<Ashok> [1] Team Comment: http://www.w3.org/Submission/2006/04/Comment

<Ashok> [2] TimBL: http://lists.w3.org/Archives/Public/www-tag/2006Oct/0061.html

<Ashok> [3] Stuart Williams: http://lists.w3.org/Archives/Public/www-tag/2008Nov/0008.html

AM: these msgs note web services use end-point references (EPRs), not URIs
... and there's a question of what if you do a GET on the URI found in an EPR
... and they [who?] has asked for more specific recommendations

<johnk> EPRs contain a URI

AM: yes, for example, the WS WGs would like the TAG to be specific about what should come back when you do a GET on the URI in an EPR

<Zakim> noah, you wanted to ask why

NM: Do we know why they're interested?

AM: there's some expectation that TAG input would help with WS specs

<Zakim> DanC, you wanted to note DaveO's suggestions that the TAG do this and ask

DanC: DaveO suggested some work nearby...

DaveO: in the Hawaii meeting timeframe, yes...

<masinter> I think working groups who want TAG feedback should try harder to express their question as an architectural issue rather than "read our specs and figure out what we should do"

DanC: DaveO, do you think the market window has passed? or is it still worthwhile?

DaveO: I've long been an advocate of filling this gap...
... I see the REST community ignoring WS-*, and vice versa...
... and then I see the OAuth community wrestling with policy problems solved in WS-policy [but which doesn't work in the OAuth context because it's not URI-based]

<masinter> i'm not convinced that this is an issue that can be resolved, REST vs. WS

<masinter> it would be good if mechanisms like authentication and naming can be common, though

DaveO: when WS-* advocates say "you can do all this automated policy negotation, provided you use SOAP messages", the REST community isn't interested.

<Ashok> How about a common mechanism to access metadata?

<masinter> well, common mechanisms might be hard, but common vocabularies for metadata would be a good first step

daveO: a couple points in particular: soap RPC into HTTP GET, bind EPRs into URIs

<masinter> general idea is: don't try to solve unresolvable problems

<Ashok> Dave, Do you have something written on this?

daveO: I can point you to work I did in that area

LMM: the perspectives of the two communities aren't clearly resolvable...

<raman> 1+ to Larry.

<raman> RESTful services are very successful, we dont need to teach them WS* religion

LMM: there are perhaps architectural approaches: common vocabularies despite different mechanisms...
... choosing one's battles is important

<johnk> +1 to the *possibility* of common vocabs

<Ashok> Raman, I'm thinking the other way around

<johnk> but tend to agree with Larry that this is still a war

LMM: I'd contrast this with the HTML situation... there's really just one HTML and we should get it right, but the world is OK having both WS-* and REST styles.

<Zakim> noah, you wanted to talk about helping the REST community discover the joys of Web Services

TVR: I agree with Larry; there's little impact the TAG can usefully have in this area [?]

NM summarizes [scribe thinks it's already recorded above]

NM: to push back a bit on LMM and TVR: the Web aspires to a level of integration that's deep and universal...
... URIs are supposed to be URIs wherever you find them [in an EPR or otherwise]

<masinter> if there's an expression of this as an architectural question, that would be helpful

NM: and there are expecations that you can do things [e.g. GET]

<raman> The TAG's history here is to traditionally have come heavily on the WS* side, mostly because (perhaps) of the bodies present. That also means that we dont get heard by the other side.

NM: [struggling to summarize tail end of what NM said]

<raman> I'm happy for the TAG to specifically answer Ashok's WSRE question.

<masinter> GET on a URI should return a representation of the resource

<raman> Bit BuBut as a continuing TAG member, I'd advice against going the route Dave Orchard suggested that the TAGshould go

<masinter> (jk)

<masinter> Would like agenda items about technical topics have an ISSUE, because I'm confused still about what they want advice about

SKW: I think it's a stretch to say they're [which they?] interested in access to metadata... they have a particular object model, and metadata is a small part of it. [?]
... I think the http://lists.w3.org/Archives/Public/www-tag/2008Nov/0008.html message is still what's worth saying. [?]

<noah> Larry: suggest we discuss the meta topic of TAG operations at F2F. This is the sort of stuff I'd like to agree on. Historyically, ISSUES have been somewhat heavyweight, with some commitment to long term TAG focus. Hence there tends to be a lot of discussion to decide whether to open an issue. That's what's happening here, I think.

NM: there's a metadata part to this; we have an action on jonathan in that area...
... then there's the specific question re WS-RA... maybe AM would like to draft something?
... meanwhile, we might want to think about whether this fits under and existing issue or should be a new one

Link maintenance in TAG documents


<trackbot> ACTION-223 -- John Kemp to attempt to fix the broken links in Mapping between URIs and Internet Media Types -- due 2009-02-12 -- PENDINGREVIEW

<trackbot> http://www.w3.org/2001/tag/group/track/actions/223

<masinter> didn't propose TAG work on this


<trackbot> ACTION-222 -- Larry Masinter to draft a note to W3C staff regarding maintaining working links, for TAG review -- due 2009-02-12 -- OPEN

<trackbot> http://www.w3.org/2001/tag/group/track/actions/222

LMM: yes, I drafted something, and I see a response from Dan noting existing policies

<masinter> I didn't want to propose the TAG should work on this, but rather we should ask the staff to handle this kind of thing in general, rather than take this up as a TAG work item

John: noting date of access is conventional for citing web documents; we could do that...

NM: isn't that implicit in the date of the finding itself?

John: note chicago manual of style etc. note date accessed

<DanC_> (I find "date of access" kinda wierd. no thanks.)

LMM: how about we ask the W3C staff to use that chicago manual of style date-of-access policy?

<masinter> hear arguments against and think they're reasonable

DanC: ISOC stable-publishes IETF drafts

<DanC_> progress on ietf.org persistence, structured archive

DanC: I recommend we use those in the future

NM: are those canonical?

<jar> Mary-Claire van Leunen. A Handbook for Scholars. = bible on citation

LMM: I'm satisfied the staff maintains a policy on links in general

<masinter> that should be W3C publication policy, not just a W3C TAG practice

<Ashok> +1

close action-222

<trackbot> ACTION-222 Draft a note to W3C staff regarding maintaining working links, for TAG review closed

NM: how about "we considered it, but on balance, didn't find it worthwile to update the finding"
... John?

<masinter> i don't agree with that

<masinter> I like jar's proposal better

<masinter> "Message to staff: We've talked about this. Please do something. [We trust you.]"

Stuart: I like the intermediate page idea...

<jar> noah: w3c ends up running a proxy site for all cited documents. not good

Stuart: I suggest that we quietly change the links to something sensible

<jar> webcitation.org

<DanC_> -1 "please do something".

<johnk> -1 also

<Stuart> +0

SKW: the staff has very tight policy for the /TR/ page, but we're not using that for findings...

NM: Dan, how about you draft something...

<DanC_> no, I'm not interested in any new policies.

<DanC_> I've done more than I think this merits already

<masinter> fix document is fine

NM: how about... ACTION: respond to the commentor

<DanC_> . ACTION: john consult with stuart and respond to the commentor re broken links in uriMediaType-9

NM: so we'll leave this to email and maybe it'll come back

close ACTION-223

<trackbot> ACTION-223 Attempt to fix the broken links in Mapping between URIs and Internet Media Types closed

Summary of Action Items

[NEW] ACTION: DanC report at March on tagSoup progress since TPAC [recorded in http://www.w3.org/2009/02/12-tagmem-minutes.html#action01]
[NEW] ACTION: jonathan summarize TAG work on metadata, with Larry [recorded in http://www.w3.org/2009/02/12-tagmem-minutes.html#action02]
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.134 (CVS log)
$Date: 2009/02/13 15:54:32 $