TAG Weekly

12 Mar 2009


See also: IRC log


Masinter, Raman, Noah Mendelsohn, Ht, DanC, John Kemp, jar, Plh
Ashok, Tim


<jar> chair: Noah

<scribe> ScribeNick: johnk

NM: canceling the call for the 19th

<DanC> (26 Mar TAG call conflicts with HTTP bis IETF meeting, which I hope to attend)

<masinter> link to agenda

pending action items

<jar> +1

NM: propose that we resolve procedures for closing open actions via email


NM: HT, can you scribe on the next call?

<masinter> s/issues/actions/

<jar> noah: "Our clocks are fine"

HT: OK...

<masinter> Note I sent email asking to not close actions that (a) have no associated ISSSUES and (b) have no other followon

<DanC> action-243?

<trackbot> ACTION-243 -- Dan Connolly to assemble minutes from SFO for day 1 based on http://www.w3.org/2009/03/03-tagmem-irc -- due 2009-03-17 -- OPEN

<trackbot> http://www.w3.org/2001/tag/group/track/actions/243

(the guilty wave hands)

NM: will schedule ongoing activity to review priorities + "big themes"

<masinter> pop

NM: any need to discus this more now?

group: no

NM: been sent a survey inquiring about TAG interest in 2-6 Sept. TPAC meeting, we should decide by the 18th



LMM: would like to understand this issue better

<masinter> we'd like to understand the proposal and possibly refine it

PLH: two groups using the same ns/media type
... can we resolve conflict by getting XHTML to stop using the ns/media type?

LMM: is there a written proposal?

PLH: no fully-fledged proposal -
... what is impact on XForms, for example?

<Zakim> DanC, you wanted to note Device Independent Authoring Language http://www.w3.org/TR/2007/WD-dial-primer-20071101/

DanC: DIAL is a way that XHTML is used

TVR: it's a profile

<masinter> what is the relationship of DIAL to the issue being raised?

DanC: It's not critical for DIAL to use the 1999 ns/media type

TVR: this is a TAG Issue because:

<DanC> the relevance of DIAL is that I'd like to know if those who recently showed interest in XHTML 2 would be happy if it were renamed "DIAL"

TVR: how to support in webarch how to discover what language you are getting at each point?

<masinter> the application/xhtml+xml media type has 'failed'

TVR: xhtml+xml has failed as a media type
... media type is thus not useful
... ns is also being abused
... 3rd thing is versioning
... Larry started discussion in HTML5 about ns as versioning mechanism
... XHTML2 went back to using the old ns
... two groups in the same space is certainly sub-optimal
... your proposal is interesting, but from webarch perspective, we will not be in any better place

PLH: agree, this doesn't resolve all the issues
... but that is not my goal

<masinter> politics intertwined with architecture

TVR: web is bigger than browser

PLH: proposal is to address one specific issue

<Zakim> noah, you wanted to make sure I understand what's intended

NM: agrees with TVRs points
... who "owns" the namespace and media type?
... we've decided to put new NS, and media type on XHTML2
... is my understanding correct?

PLH: yes

<masinter> need to come up with agreement on what the issues are and what the choices really are, put into a form where the membership can really review and comment on it. as currently framed, the "idea" is incoherent and disruptive

PLH: new namespace solves the bootstrap problem

<masinter> different architectural choices in this space will impact companies and their businesses in different ways, and consensus will be hard because of the economic impacts

TVR: what I hear is we create YA media type

<DanC> ("that won't work" ... umm... really? as far as I understand the DIAL use cases, they don't need MS IE to grok a new media type)

<DanC> larry, when you verbally ack somebody, it helps to simultaneously "ack next"; Zakim is smart enough to unmute people who get the floor

<Zakim> ht, you wanted to discuss XForms

CORRECTION: by using the OLD namespace in XHTML2, you solve the bootstrapping problem by using JS to handle the XHTML

HT: from TAG perspective, NS competition is place where we should start

<Zakim> masinter, you wanted to suggest we work with PLH to make the choices clearer

LMM: clear that is important for w3c, decision is not speculative or forward-looking
... choices for members are not clear
... cannot see it being possible to make an informed choice on a proposal that is not fully-fledged
... would be a good service to the community to write out the issues and choices
... TAG should take on such a work item

<Zakim> raman, you wanted to add that defining Web purely as a browser for N years after having defined it *without* the browser fo r8 years is a mistake. We need a balance

TVR: risk is that you equate the browser with the Web

<DanC> (good point; the risk that "future of browser" will be understood as "future of W3C" is pretty high. not a bet I'd take)

TVR: problem is that we have made the mistake of ignoring the browser, and don't wish to swing back too far the other way

PLH: yes, it is a difficult balance
... we will have a panel at the AC meeting

(scribe missed most of the names)

<noah> PLH: Steven Pemberton will be at the AC meeting and on the panel.

<masinter> people come to the AC meetings to represent their companies, and need to review their company's point of view. That point of view needs to be based on choices that have been explained clearly. The current proposal isn't.

TVR: DAISY has been based on XHTML2

<jar> http://www.daisy.org/ ?

http://www.daisy.org/, yes

<noah> Larry, I'd be curious to know what the specific goal would be of TAG work (not unsympathetic, by the way)

<ht> I will now be at the AC, with my Edinburgh hat on

<masinter> create a background document which lays out the issues and considerations we've already discussed, and the relevant sections of AWWW and the impact it would have

<Zakim> noah, you wanted to discuss content creators

TVR: back to language design issues

<DanC> oops; sorry

<masinter> i'm using the queue but giving raman slack

TVR: web has always been extensible (netscape plugins eg)
... then JS+DHTML

<jar> yay greasemonkey!

TVR: then greasemonkey

<DanC> (hmm... stack: plug-in API, browser extension, greasemonkey)

<masinter> different kinds of extensibility: protocol, MIME type, namespace, ....

TVR: can extend on the protocol handler, or on the MIME type, or on namespaces

<noah> Seeing the list plugin, extension, greasemonkey all together for some reason reminds me that at the F2F we noodled on the TAG focusing more on Web security.

TVR: believe this is a big mistake

DanC: what should we do instead?

PLH: I'm not addressing the general problem with this proposal

TVR: not concerned about the specific item per se

LMM: panel should try and frame a TAG overview of the arch issues, pointing to relevant findings and issues

<noah> Larry, should we give someone an action to prepare a TAG position for either you or some other TAG member to present?

LMM: I can't be on the panel as both TAG rep and company rep

<DanC> (then LMM said he *can* do both of those. hmm.)

LMM: concerned that this topic has this proposal already as if there is some reasonable resolution already
... would like to frame the issues coherently first

TVR: yes, feels like a fait accompli

<Zakim> masinter, you wanted to ask TAG members if they're willing to take on this issue and immediate actions

<Zakim> noah, you wanted to discuss content creators

NM: if there's some chance that we need to present a TAG position, we need to work on this soon
... should we assign an action to guide TAG through such a process?
... a possible issue about who represents TAG on this panel
... wanted to mention the content creators

<masinter> I didn't volunteer to be on the panel to represent either Adobe or the TAG but to add what I hoped was an informed perspective

NM: if we can have fair representation from content creators in this discussion, that would be good

<Zakim> DanC, you wanted to note that MS implements a namespace hook

DanC: MS has implemented ns support (URI-based extensibility) since IE6

<Zakim> ht, you wanted to disagree with TV about NS and media type dispatching

<masinter> and representation can't really happen until we've discussed the issues and come to a conclusion, and I don't think we'll have a fully-fledged "considered position". I'm not proposing that the TAG have a "considered position" but that we document the analysis of the issues we've done so far in order to make the discussion at AC more productive

HT: agree with Noah that we continue to struggle to find content-creators willing to talk about this issue
... when NS + media-type extensibility is made available to developers, it *is* used
... using HTML+SVG because MathML works now
... this wasn't possible a year ago
... that entry point into extensibility is worth fighting for

<noah> Henry reminds me that more specialized communities, like scientists, may have an important perspective on extensibility.

HT: takes us back to the core question: are we prepared to declare text/html media-type and 1999.xhtml ns an "open" framework?
... starting the discussion at the procedural level is a mistake

<masinter> i say 'versioning' in a named or branched sense, not just a linear way

(agreement that browsers have NOT written off NS-based extensibility)

<masinter> it sounds like the proposal is based on false assumptions?

<masinter> HTML+SVG is being discussed

<DanC> (I'm persuaded the current proposal is misleading)

<masinter> want to push again on the framework of the panel and the AC meeting. The "panel" should be an explanation of the issues and the point of view, and not an uninformed debate

<noah> BTW: if you look at materials XML schema developed in exploring versioning they strongly support what Larry is saying about the distinction between namespaces and language versions

LMM: if you don't use NS for versioning, then you can use some other indicator (eg. DOCTYPE or explicit version number attribute)
... should push on this notion of versioning being a linear progression

<noah> As I wrote in the TAG blog article, I think you only need to distinguish versions in the representation IF the same content would have different meaning in two or more versions. If all that happens is that content becomes legal or becomes illegal, you can tell that without labeling the versions at all

<noah> See blog entry at: http://www.w3.org/QA/2007/12/version_identifiers_reconsider.html

LMM: different WGs are talking past each other - don't want to see a panel as a debate
... panel should be a discussion without a presumption that there will be some immediate vote or action

<ht> HST endorses LMM's analysis: see http://www.pdfpower.com/XML2005Proceedings/ship/82/XML_2005_82.HTML

LMM: TAG has worked on versioning extensively, and should use this chance to solicit opinion from members to create a coherent proposal

<raman> q, ack

<Zakim> masinter, you wanted to correct Noah's mischaracterization of what I said and to point out that it is possible to have two different languages with the same namespace, as long as

<Zakim> raman, you wanted to add that Larry is well-respected as Larry --- I wouldn't split hairs about Tag concensus position and Larry being the message bearer

<masinter> i'm really sorry, need to get off phone for a bit

NM: is it really only in HTML5 discussions that are saying "look, if I do SVG/MathML, I want to do it without ns?"

HT: yes, my understaning

TVR: (agrees)

<masinter> can stay on IRC

<masinter> sorry

PLH: panel is not there for a debate or to make a decision
... but to get feedback

<DanC> (I don't know how to answer noah's question; many of the loudest voices in the HTML 5 discussion represent browsers)

NH: is there anything that the TAG can help with in support of this panel?

PLH: if TAG was willing to write something about this issue, that would be helpful

NM: this is a complex topic which includes opinions which do not yet converge
... (a little surprised at this issue taking the turn it has towards TAG focusing on versioning vs. other HTML problems)
... (repeats question to PLH about what specifically the TAG could do)
... if we should do anything more than we're already doing, we need to decide that now

<raman> 1+ to HT presenting on behalf of the TAG

<noah> Yes, I like that. PLH, is there room? Is that a good thing?

HT: point that Larry raised - the media-type to namespace to formal version identifier should be made on the panel

<plh> there is room, and I welcome Henry

<noah> Great. Let me set up the necessary actions when Henry is done talking.

<raman> at this point, even getting a one-level indirection/extension mechanism would be a win. With 3 possibilities, all 3 are abused, and people need to invent itunes:

NM: HT, can you circulate a proposal?

HT: accepts

<DanC> (I don't need to be in the critical path; I'm happy for HT and LMM to say what they like, whether on behalf of me, the TAG, or otherwise.)

<scribe> ACTION: ht to circulate a proposal framing this issue [recorded in http://www.w3.org/2009/03/12-tagmem-irc]

<trackbot> Created ACTION-246 - Circulate a proposal framing this issue [on Henry S. Thompson - due 2009-03-19].

<masinter> will help, think this is quite valuable


NM: JAR took ACTION-227

<DanC> action-227?

<trackbot> ACTION-227 -- Jonathan Rees to summarize TAG work on metadata, with Larry -- due 2009-02-24 -- PENDINGREVIEW

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

NM: proposes to close 227


close ACTION-227

<trackbot> ACTION-227 Summarize TAG work on metadata, with Larry closed


<DanC> action-227: http://www.w3.org/2001/tag/2009/02/metadata-survey.html

<trackbot> ACTION-227 Summarize TAG work on metadata, with Larry notes added

<masinter> sorry can't get back on phone

<noah> No problem, Larry.

<noah> Thanks for getting back to us.

DanC: would like to wait for Larry on this item...

proposal to adjourn

<jar> Discussion tabled pending Larry's participation.


<masinter> sorry

<masinter> i'll check in with HT

Summary of Action Items

[NEW] ACTION: ht to circulate a proposal framing this issue [recorded in http://www.w3.org/2009/03/12-tagmem-irc]
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.135 (CVS log)
$Date: 2009/03/18 10:06:57 $