W3C Technical Architecture Group (TAG) - Teleconference of 11 June 2007

11 Jun 2007


See also: IRC log


Tim Berners-Lee, Dan Connolly, Noah Mendelsohn, Henry Thompson, Rhys Lewis, T.V. Raman (2nd half of meeting), Norm Walsh
Rhys Lewis, David Orchard
Stuart Williams
Noah Mendelsohn


Review of Agenda

SW: Any proposals for agenda changes?


Future meetings

SW: Teleconference next week on the 18th. Norm, can you scribe?

NW: Unfortunately, no, but the following week I can.

SW: Thank you. Henry, you're the next of those here today. Can you do next week?

HT: Yes, I can do it next week.

<timbl> XSLT LRES?

Review of minutes from 30 May - 1 June 2007 F2F

<DanC> Cc: tag@w3.org

<DanC> Subject: Re: Draft minutes from 1st June 2007

<DanC> Date: Mon, 11 Jun 2007 10:51:20 -0500

DC: We should have formal actions recorded for some of the tasks we agreed to take on.

NM: More appropriate to make them formal this week rather than claiming retroactively we did so last week.

<Stuart> Actions volunteered at F2F:

<Stuart> Noah to revise Self-Describing Web finding

<scribe> ACTION: Noah to revise Self-Describing Web finding [recorded in http://www.w3.org/2007/06/11-tagmem-irc]

<Stuart> Henry to revise URNsAndRegistries-50 finding

<scribe> ACTION: Henry to revise URNsAndRegistries-50 finding [recorded in http://www.w3.org/2007/06/11-tagmem-irc]

<Stuart> Rhys to revise Dereferencing HTTP URIs finding.

<scribe> ACTION: Rhys to revise Dereferencing HTTP URIs finding. [recorded in http://www.w3.org/2007/06/11-tagmem-irc]

<Stuart> Dave Orchard to revise Versioning Findings.

<scribe> ACTION: Dave Orchard to revise Versioning Findings. [recorded in http://www.w3.org/2007/06/11-tagmem-irc]

RESOLUTION: The minutes of the 30 May - 1 June 2007 F2F meeting as linked from http://www.w3.org/2001/tag/2007/05/29-agenda are approved.

<DanC> 2007/06/11 14:35:38 http://www.w3.org/2001/tag/2007/05/30-minutes 2007/06/11 16:20:53 http://www.w3.org/2001/tag/2007/05/31-tagmem-minutes 2007/06/11 15:37:39 http://www.w3.org/2001/tag/2007/06/01-minutes

F2F Planning

NM: It appears we can get a room to meet on Wed. the 19th of Sept. at IBM Hursley.

SW: I propose we meet until mid-afternoon on the Wed. Know that TimBl cannot join us.

TBL: Easier there than Southampton?

SW: Thought you didn't have a room.

TBL: No, room can probably be had.

NM: From IBM's point of view, we can hold off for quite awhile on committing to Hursley vs. Southampton. Big question is, would the choice affect anyone's answer?

TBL: How late in the afternoon?

HT: Flights to US tend leave by noon anyway. Therefore, only pertinent constraint is UK folks. If we aim mid-afternoon, then UK folks can get home.

SW: 3 PM?

NM: Fine with me, as long as we get at least 3/4 day, and don't lose anyone else.

SW: Fine with me.
... Propose that we meet 2 3/4 days in the UK.

RESOLUTION: The UK TAG meeting will be 2 3/4 days in length.

<DanC> (I need the full dates and the location to the nearest airport in the record... trying to remember...)

<DanC> Southampton 17-19 Sep 2007

SW: I'll check on Southampton. Noah, OK if we use Hursley as fallback?

NM: Yes, don't need to know until sometime in the summer.

SW: Closest airports are Southampton (hard to get flights), Heathrow & Gatwick.

DC: Who's hosting?

SW: Tim, in his professorial role at Southampton.

<timbl> That doesn't mean access to large amount of dinner funding.

<timbl> But it does guarantee lots of restaurant recommendations :)

Blog and Tracker

<Stuart> http://lists.w3.org/Archives/Member/tag/2007Jun/0000.html

SW: Dan's done some preliminary research. What do we need to do to get going? See: http://lists.w3.org/Archives/Member/tag/2007Jun/0000.html .

DC: You've got to act interested if you want me to do work. I didn't get any responses.

<Stuart> http://www.w3.org//2005/06/blog/skins/sweo/

NW: You seemed to need graphics designers, and that's not me.

<Stuart> http://www.w3.org/2005/06/blog/skins/sweo/

<DanC> I think the sweo blog is http://www.w3.org/2001/sw/sweo/

DC: If we use sample skin, I suppose we don't have graphics work to do.

<ht> That page works for me

DC: Can fix later but, first impressions matter.

NM: Functional but bland. OK if the content is compelling, but the design won't greatly appeal to those who weren't already intrigued.

TBL: Didn't immediately jump out as "blog"

TVR: Lead with our strength, and graphics appears not to be it. Suggest we just make the content good.


TVR: No dates please.

<DanC> TV suggests http://www.w3.org/tag/blog/

NW: 2001/tag/blog?

TVR: Please no.
... http://www.w3.org/tag/blog/

SW: Blog == TAG Home Page?

NM: Prefer not. Make a good home page, and have it link the blog. Maybe pull a feed of a few headlines into the home page if we think it's worthwhile.

SW: Can we make this happen?

???: Tim as webmaster in chief can make it happen.

TBL: Anyone object to 2001?

TVR: I do.

<DanC> (I can accept /2001/tag/ , but not HT's rationale. This blog is all about not perpetuating old sins.)

HT: We made the 2001 mistake awhile ago.

NW: Yes, there will be lots of downstream questions about which future content should be under 2001 vs not.

TBL: My concern is that in 2050, there could be many, many different things called "tag".
... For awhile, the top level of the W3C site was full of lots of words of the English language.

TVR: As 2001 goes further and further into the past, readers who check the URI will tend to think that even new findings named that way must be very old.

SW: Propose we use 2001/tag/blog. Objections?

TVR: I object.

Others are silent.

DC: Abstain.

SW: I will suggest we resolve, in spite of objection.

RESOLUTION: The URI of the TAG's blog is to be http://www.w3.org/2001/tag/blog


SW: We received a request from Leo Sauermann at http://lists.w3.org/Archives/Public/www-tag/2007May/0086.html to review some of their work http://www.dfki.uni-kl.de/~sauermann/2006/11/cooluris/ . I am unsure how to proceed. Is it appropriate to do such a review?

NW: We can, don't have to. I think in this case it's an OK thing to do.

DC: Rhys not here? Too bad.
... I took a look, and informally, it looks like a lot of it is OK.
... Might even be interested in asking whether it's a basis for the finding Rhys is doing.

HT: Aims and audience seems different.
... I think we can respond: "this is basically fine, some details need tuning to be quite right".

DC: Why is audience different?

HT: Well, maybe I mean the level of the intended audience might be a bit different.

DC: Hmm. At the F2F, I thought we were helping people to choose URIs, and that's what this thing seems to be doing.

HT: You said that was your goal at the F2F, but I'm not sure that historically was quite the scope.

TBL: I think you suggested a title change, and people didn't agree.

DC: I asked what our goal was, and I thought that's what I heard.

TBL: Not from me. I don't want to drop treatment of the other important stuff. Discussing choice of URIs is fine, but not as an alternative to the other material.

SW: Tim, what do you think the finding is focussing on?

TBL: What would I like it to cover?

SW: Yes.

TBL: I want it to formalize what happens in HTTP, so readers understand the implications of setting up a server in certain ways. It should explain the related issues of choosing the URIs, and it should cover the semantics of 301, 302 and 303, and also the semantics of # in Web architecture, leading to discussion of how to assign semweb uris.

DC: I can't imagine people wanting to know that.

HT: I do. Our existing work discusses info resources, etc., and this tells the rest of that story.

DC: Turning it upside down is so much more approachable.

TBL: Well, the title could be either way.
... Could still be "how to choose URIs", but I'd be nervous about that if the title were taken to narrow the scope.
... I think the TAG's goal is to connect everyday experience to the specs.

<Noah_> Hear, Hear!!

DC: We're discussing editorial issue, but the editor's not here.

<timbl> TBL: ... But also to connecte it to the architecture explained. To explain why.

<timbl> To explain how it works, an why those choices must be made for URIs.

SW: Shall we respond to Leo that we appreciate their writings, but need to focus on our own work first?

NM: I could easily support being more helpful to them, but I confess I don't have the time myself in coming weeks. We could point out what's good, what seems to need work, and where we don't yet know the relationship to our own upcoming findings.

SW: I can review.

NW: I can too.

SW: tag or www-tag lists?

NM: Don't care. I prefer doing this things in public, but sometimes it's better to take a round on the private list first if you think earliest drafts would be too rough and therefore confusing.

<scribe> ACTION: Norm and Stuart to review http://www.dfki.uni-kl.de/~sauermann/2006/11/cooluris/ [recorded in http://www.w3.org/2007/06/11-tagmem-irc]


<ht> http://www.w3.org/2001/tag/doc/elabInfoset.html

Original document: http://www.w3.org/2001/tag/doc/elabInfoset.html

HT: That's nearly 6 months old.

<Stuart> Also henry's message from today http://www.w3.org/mid/f5by7iq34o1.fsf@hildegard.inf.ed.ac.uk

<ht> http://lists.w3.org/Archives/Public/www-tag/2007Jun/0033.html

HT: We discussed in Boston in Feb, but stalled when we discovered that Tim and I (at least) had differing views of the granularity of the interaction of 1) construction of elaborated infoset with 2) interpretation of XML docs as a whole.
... I discussed a notional 2nd step, I.e. after parsing, but Tim wasn't happy with that. He felt that it all had to be better integrated with compositional semantics for XML in general. I prepared the email at http://lists.w3.org/Archives/Public/www-tag/2007Jun/0033.html for our recent F2F, but we didn't get to it.
... Question: how do we talk about balance of control for particular XML vocabularies vs. generic mechanisms such as XInclude? It seems to come down to a question of quoting. By quoting, we mean specifically, some part of a language has the implication: don't do elaboration
... Question: should way say that the way forward is to say that particular specs can say "this spec operates not on a traditional infoset resulting form a parse, but from the elaborated infoset"? However, we also need a way for, say, the GRDDL spec to say something like "the RDF namespace is a quoting namespace with respect to elaboration"

SW: Tim?

TBL: Is this about the signals?

HT: No, that's Noah's (pending) question. This about having a given specification say of its input: elaboration stops when ....

TBL: But, we need a story about lanuages that don't say anything.

HT: Opt in or opt out? Default for old specs is you don't elaborate, but a new spec can opt in.

TBL: I'm OK with a spec saying "You should elaborate everything >except< this..."

<DanC> transparent is elaborating, briefly.

TBL: Thinking about RDF: "Anything which has a parsetype attribute which is an XMLliteral is opaque (non elaborating), otherwise transparent (elaborating)"

HT: Stipulate that this work were published as a full Recommendation, with all the corresponding reviews. It would have the consequence that we are grandfathering in XInclude and perhaps decryption as if they had said from the start: "we are part of standard elaboration". Or maybe we need to republish those recs? Not sure.

<DanC> (hmm... maybe, HT. I'm not juggling too many things in my head.)

<DanC> (now)

TBL: The advantage here is that there's a standard form of elaboration to which you can point.

HT: Agreed. There are two types of opt in: (1) Type you would do with XML schema 1.1 vs. (2) some putative new general purpose elaborating technology, XMLExclude perhaps. No clue what it would do, but it's clearly at the level of XInclude.

<Zakim> Noah, you wanted to ask about quoting >namespaces<. Don't think it's always namespaces.

<Zakim> timbl, you wanted to suggest that toadd new forms of elaboration you EITHER mark your element as opaque (not automatically elaborating), OR you add to the set of functions.

NM: Having elaborating "namespaces" is a common idiom, but it's elements that have semantics. We should avoid talking about elaborating ns.

HT: Agree.

TBL: Me too, but I was on the queue to talk about something else:
... Consider two cases. In case one, there's strange ad hoc processing going on, and the results are unlikely to be useful for recursive interpretation. Struggling for examples. Perhaps XSL template-like thing.
... Other example is a new general transform function.
... In the first case, you just note it as opaque. In the second, you need to rev the interpretation of what it means to be (future version of) xmlFunctions-capable to include the new transform.

HT: Exactly what I had in mind.
... Let me leave it at that, but I haven't yet dealt with Dan's concern. Will come back to that.

<Zakim> Stuart, you wanted to ask how much of the context for the quotations needs to propagate with the quotation - eg ns bindings

SW: Is there an issue of asking how document contexts, e.g. namespace bindings, propagate with quoting?

HT: I think the person responsible for processing a quoted region can restart elaboration.

<timbl> Yes.

HT: XInclude for instance, which has story higher up, has a story about namespace fixup.

<DanC> (HT, my concern might be addresses. I think I need a quiet room (with your new draft) to be sure.)

SW: Henry, do you have what you need to keep going?

HT: Yes, it's worth writing a new draft.

<scribe> ACTION: Henry to prepare new draft of xmlFunctions-34 by mid-July. [recorded in http://www.w3.org/2007/06/11-tagmem-irc]

Passwords in the clear (unscheduled topic)

TVR: At the Web conference, Mary Ellen Zurko suggested she might be willing to contribute text on passwords in the clear.

SW: We'll discuss next week, and I'll invite Ed Rice.
... We are adjourned.

<DanC> http://dig.csail.mit.edu/issues/tabulator/issue?@columns=title,topic,id,activity,status,assignedto&@sort=priority&@group=assignedto&@filter=status&@pagesize=50&@startwith=0&status=-1,1,2,3,4,5,6,7&@dispname=Developer%20Status

Summary of Action Items

[NEW] ACTION: Dave Orchard to revise Versioning Findings. [recorded in http://www.w3.org/2007/06/11-tagmem-irc]
[NEW] ACTION: Henry to prepare new draft of xmlFunctions-34 by mid-July. [recorded in http://www.w3.org/2007/06/11-tagmem-irc]
[NEW] ACTION: Henry to revise URNsAndRegistries-50 finding [recorded in http://www.w3.org/2007/06/11-tagmem-irc]
[NEW] ACTION: Noah to revise Self-Describing Web finding [recorded in http://www.w3.org/2007/06/11-tagmem-irc]
[NEW] ACTION: Norm and Stuart to review http://www.dfki.uni-kl.de/~sauermann/2006/11/cooluris/ [recorded in http://www.w3.org/2007/06/11-tagmem-irc]
[NEW] ACTION: Rhys to revise Dereferencing HTTP URIs finding. [recorded in http://www.w3.org/2007/06/11-tagmem-irc]
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.127 (CVS log)
$Date: 2007/06/11 23:46:49 $