Approved minutes of TAG Teleconference of 22 January 2009

On its teleconference of 12 February 2009 [1], the TAG approved the 
publication of minutes of its teleconference of 22 January 2009 [2].  I 
see no record of these being announced on this list, so I am doing it now. 
 A text version of the minutes is attached.  Thank you.



Noah Mendelsohn 
IBM Corporation
One Rogers Street
Cambridge, MA 02142



                               - DRAFT -

                              TAG Weekly

22 Jan 2009



   See also: [3]IRC log



          Tim Berners-Lee, Dan Connolly, John Kemp, Ashok Malhotra,
          Larry Masinter, Noah Mendelsohn, David Orchard, T.V. Raman,
          Jonathan Rees, Henry S. Thompson, Norm Walsh, Stuart Williams

          Stuart Williams

          David Orchard


     * [4]Topics
         1. [5]minutes
         2. [6]Issue 36 site (meta) data
         3. [7]self-describing web 61
         4. [8]widgets, web apps wg
         5. [9]WebApplications
     * [10]Summary of Action Items


   <Stuart> [11]


   minutes of 15 Jan 2009 approved

   Henry will scribe next week

Issue 36 site (meta) data

   jar: I've looked at sitemeta and have a few comments



   <jrees> Eran's note:


   zakim next agendum

   jar: discovery has 3 components: header, link header, link
   ... it's out of scope what the descriptor is
   ... the rationale section was good and well justified.

   <DanC> (I find "Appendix A.2. Analysis")

   jar: small comments, like resource discovery -> resource descriptor
   ... what if you get different metadata through different methods.



   jar: says you must produce the same description and location.

   <DanC> (jar's comments, or at least some of them:
   [15] )


   jar: a large class of additional languages might want an html link

   tbl: presumably you could put the html:link in another language.

   jar: I agree

   johnkemp: you can use any link header as long as it uses the link
   ... page 7 hammer-discovery-00

   <masinter> the computation of the "base" of relative URIs is generic

   <timbl> sorry pringingthe dco off

   <timbl> doc

   <masinter> style cascades

   tbl: user can edit html doc, and would want multiple link headers

   <masinter> someone is drying their hair

   <masinter> much better

   <Zakim> DanC, you wanted to ask for a story involving real-world
   parties on the critcal path to deployment

   <timbl> Sorry, for the noise

   <Ashok> In an example related to the identity space, an end-user can
   use a URI as an identifier for signing into web services, and in
   turn, the web service can discover more information about the user's
   resources and preferences such as who did the user delegate their
   identity management to, where they keep their address book or list
   of social network friends, where their profile information is stored
   to reduce signup registration requirements, an

   <timbl> I would suggest as an alternative: if the rtwo resource
   descriptors found though different paths of the 3 diffferent wy have
   the same information, then they should have the same URI to save the
   client looking up both.

   <DanC> (still too many pronouns)

   <Zakim> jrees, you wanted to answer dan with story

   <DanC> (the canonical openid story is: I want to comment on a blog
   item that I just read, and rather than making a new
   username/password, I just use my openid)

   jar: looked at openid in wikipedia. goal is to reduce the number of
   registrations you provide. At a new site is I use my magic string.
   ... I still have to authenticate, but I use one string across the

   <DanC> (ecommerce is a bit of a stretch, given security properties
   of OpenID, but I'll stipulate for now)

   <ht> Am I the only one who is confused as to why we are talking
   about OpenID?

   jar: we're talking openid because Dan wanted 1 story.
   ... rather than coming up with a new namespace, perhaps a person
   could provide one string like a home page or mailto: uri.

   <DanC> (using openid to find a public key... gee... you're not going
   to trust that connection, are you? )

   jar: this can be used to find resource descriptors..

   <masinter> there are no "name services" there are only "name lookup
   services". It's a common game to talk about "assigning names" when
   the truth is "promising service for name lookup"

   I'll note there is a canonical mailto: URI to http URIs for openids.

   masinter: people confuse name assignment with name lookup
   ... same problems with dns, oids, ....
   ... what people are really getting is a promise that there will be a
   server that can be used to find information.
   ... need to be clear about future service vs the name
   ... because the identity contains the scheme for how to get
   information about the service (http:).

   <DanC> (I find it hard to get motivated to look into the technical
   details until we know more about the market and deployment

   masinter: Now we are asking what's the way to get the metadata about
   the service.
   ... There can be many different locations for the link or relations
   to the metadata about the service
   ... seems more likely that the different sources may have different

   <Stuart> +1 to what Larry just said

   tbl: do you mean different but not conflicting?
   ... this is all from the same site, the same dns name.
   ... stuff that comes demonstrably from the same source as the doc
   ... different sites may have different parts under different control
   ... they are different information

   masinter: there is a spectrum from static to dynamic. for example
   rating is static

   tbl: could also have something like subscription expiry

   masinter: I wouldn't expect the spec to disallow conflicting
   information because it's likely to happen

   ashok: different users have different access to resources

   <timbl> That's true: Anotehr reason for having different metdata
   reource with different content is they may be accessible o diff
   categories of user.

   ashok: metadata spec has an example of where the metadata is inside
   the representation.

   <Zakim> jrees, you wanted to suggest that probably Eran wants to
   permit discoverers to stop when they find the first resource
   descriptor - in this protocol.

   jar: the specs are trying to do 1 thing and do it well.
   ... obviously people can get metadata from everywhere.
   ... this gives a way to communicate information about the name of
   the resource descriptor
   ... is this sufficient such as different mechanisms? no. This
   doesn't work for all different file types? no.
   ... to ask it to look into pdf is making it too complicated

   ashok: both specs talk about revised versions coming out. Let's take
   it up again.

   <Stuart> close action-202

   <trackbot> ACTION-202 Review
   [16] closed


   <masinter> suggest a more definitive action, e.g., to encourage W3C
   review of these drafts, or to track the discussion on some mailing
   list, or to start a new mailing list about it?

   <DanC> issue-36: note


   <trackbot> ISSUE-36 Web site metadata improving on robots.txt,
   w3c/p3p and favicon etc. notes added

   <DanC> issue-36: <jrees> Eran's note:


   <trackbot> ISSUE-36 Web site metadata improving on robots.txt,
   w3c/p3p and favicon etc. notes added

   <DanC> ("the mailing list that's talking about it" is which? I think
   it's www-talk for nottingham-site-meta , but I dunno for
   hammer-discovery )

   skw: there's been a general interest in web community, and semantic
   web, about this area.

   (didn't eran set up a mailist list for the discovery??)

   <DanC> (dunno, dorchard ; haven't followed eran's work much at all,

   there's a discovery mailing list:

   on the discovery feedback, Eran said "Feedback would be highly
   appreciated either directly to me or on the mailing

   <DanC> ah. good.

   skw: jar's next steps?

   jar: comments to Eran, keep following
   ... I'm in conversation with lisa Dusseault(sp?) and mnot about
   303s. very interesting

   <masinter> Lisa Dusseault

   <masinter> Dusseault

   <masinter> sorry

self-describing web 61

   <DanC> (I didn't manage to propose anything re squatting nor
   x-tokens, re ACTION-211 ; I'm OK to withdraw if there's critical
   mass of support to publish the current draft)

   mark baker produced a branch draft, henry has commented on that

   skw: have your comments been addressed henry?

   ht: I don't think Noah has done the cleanup on the (g?)rddl example.
   ... should only take 30 seconds to fix the grddl example.

   <noah_waltham> No, Mark did not produce a branch draft. I did, in
   response to concern from Mark. Henry has since suggested further
   changes, which are fine with me.

   ht: wrt noahs? (should be mark bakers?) I agree with the first part
   but not the new paragraph.

   <noah_waltham> Yes, I have agreed to do GRDDL cleanup, but it's not
   done yet.

   <noah_waltham> No, I'm not aware that Mark did a draft. I did.

   <DanC> (ht's alternative, I suppose:
   [19] )


   <noah_waltham> Henry is accepting first sentence of my branch draft
   change, replacing the larger change at the end of 4.2.3. Again, I am
   agreeable to Henry's proposed change.

   ht: if you are concerned with the xslt example, the problem is the
   root element is not in the xslt namespace.

   <DanC> DanC: the XSLT literal result issue isn't addressed, but
   maybe I'm OK to publish anyway

   <masinter> RFC 3023 says " The W3C, and the W3C XML Core Working
   Group, have change control

   <masinter> over the XML specification."

   ht: there is an issue we haven't addressed which is that comments on
   self description wrt a live full blown xml document and don't treat
   xml documents (scribe loses the thread)

   <timbl> Not as themselves?

   dc: something might look like a check element but it's not

   <masinter> Well, I was puzzled by the characterization of the
   "issue" was over what RFC 3023 says. The issue is what the XML
   specification says (or rather, doesn't say)

   ht: xslt ought to have a media type for when xslt is embedded inside

   dc: self-describing web breaks down when....

   <DanC> Vendor publiseshes po-template.xml

   dc: vendor co publishes po template.xml

   <DanC> as application/xml

   <DanC> and the root element bears [20]http://vencorco.example/po

     [20] http://vencorco.example/po

   <DanC> and below the root it has XSLT goo

   this is the same problem with documents that contain xinclude and is
   the generic xml processing model problem.


   vendorco has a place for extensions including xslt and example line

   ht: purchase order contains minimum 4 tables, so template contains 4
   ... xslt wg didn't have a leg to stand on when they said it's not a
   purchase order

   <masinter> the nature of the conversation confuses me; it's as if
   the TAG is trying to decide what (ambiguous) documents meant based
   on what they said, rather than just trying to update them to be
   clearer and what they *should* say

   <jrees> danc: I don't think it's a critical bug to fix, but it would
   be nice to have a paragraph about it

   <noah_waltham> Please keep in mind that, while the finding surely
   has to be accurate, the point is to teach about self-description and
   URI-based extensibility, not to dive deep on nuances of 3023 and

   <jrees> danc: content with no change

   <Norm> The salient point here being the observation that XSLT
   template documents must have an xsl:version attribute on the root

   danc: content because the root element must have the xslt version

   <noah_waltham> Also: given that I won't be on the call for a few
   mins, please keep in mind that I will be very dependent on the
   minutes being clear as to what, if any changes have been "agreed"

   <DanC> +1 new para in


   <jrees> there was a missing " ... i sent email to www-tag just now

   <ht> Propose to ask the editor to publish after 1)

   <ht> Fixing the GRDDL example;

   <ht> 2) Adopting my proposed final para for 4.3.2

   <DanC> +1 ammendment from jar

   <DanC> +1 ammendment from jar


   <jrees> 3) change the sentence containing "owl:sameAs" so that it
   uses an OWL-DL example (see email from jar to noah)

   <jrees> i want to give noah lots of freedom here

   noah: I want less freedom not more. I'd like specific wordings

   <jrees> ok, for less freedom, just use one of my suggestions.

   jar: as proposed

   dc: as proposed

   <noah_waltham> The branch draft URI


   noah: as proposed = with 3 changes?

   <jrees> yes

   noah: fine any way, as proposed is fine

   norm: fine

   skw: fine with 3 changes

   tbl: concur

   dc: tbl, please offer informed opinion

   <DanC> (one of my hesitations is that I don't know what the "GRDDL
   example" is)

   <masinter> if the question is whether to publish the document as a
   finding: my opinion is that I would likely have lots of comments on
   the document and raise numerous issues and you should publish it
   quickly so I don't have a chance to do that.

   <noah_waltham> Proposal: we will publish SDW as proposed in
   with 3 changes 1) last para of 4.2.3 as proposed by Henry 2) GRDDL
   example vcard reference 3) one of Jonathan's proposed changes for


   <jrees> noah, (as i said in recent email) please add missing " in
   5.2 "[25]> ... i don't
   think anyone will object


   <noah_waltham> Noah will review final version with Henry.

   <noah_waltham> Proposal: we will publish SDW as proposed in
   with 3 changes 1) last para of 4.2.3 as proposed by Henry 2) GRDDL
   example vcard reference 3) one of Jonathan's proposed changes for


   skw: calling the question on noah's proposal

   <noah_waltham> Proposal: we will publish SDW as proposed in
   with 3 changes 1) last para of 4.2.3 as proposed by Henry 2) GRDDL
   example vcard reference 3) one of Jonathan's proposed changes for
   OWL sameAs


   <DanC> Jan 19 23:03:23 2009 UTC is the date of


   skw: any objections?

   tbl: worried about the diagram?
   ... pathological the same.

   noah: I think I got some help from Norm on the diagram sizing.

   tbl: it's been made to fit but it doesn't quite work with the
   ?function de-ref?

   noah: document denotes URI should go down

   <DanC> (I can't read the diagram and wouldn't miss it if were gone,

   skw: I think I'm hearing an objection from tbl.

   <noah_waltham> Proposal: we will publish SDW as proposed in
   with 3 changes 1) last para of 4.2.3 as proposed by Henry 2) GRDDL
   example vcard reference 3) one of Jonathan's proposed changes for
   OWL sameAs 4) rearrange diagram so dref is clearly the input


   <noah_waltham> Norm: I will work on it.

   <DanC> +1 Jan 19 23:03:23 2009 UTC draft + 4 changes

   <ht> HST has to leave in a minute, I'm happy with the way this is

   <timbl> I can review it tomorrow at 10:00 ET

   <noah_waltham> Norm: can you have a diagram by 1);00 ET tomorrow?

   skw: calling the question on the latest proposal?

   <noah_waltham> Norm: can you have a diagram by 10:00 ET tomorrow?

   no objections

   no abstentions

   woot woot!

   noah: thanks for the support over the past 18 months

   <DanC> (I count the critical path at: norm, noah, ht)

   no objections from TAG members

   no abstentions from TAG members

   <DanC> (the due part goes at the end)

   <noah_waltham> ACTION: Noah to publish SDW finding, with 4 changes
   as noted in minutes of 22 Jan 2009 tag telcon due 29 jan 2009
   [recorded in

   <trackbot> Created ACTION-216 - Publish SDW finding, with 4 changes
   as noted in minutes of 22 Jan 2009 tag telcon due 29 jan 2009 [on
   Noah Mendelsohn - due 2009-01-29].

widgets, web apps wg

   <DanC> "Cross-Origin Resource Sharing"


   <timbl> Norm, get it in the middle

   <timbl> 'deref' is the entry point

   <DanC> "Cross-Origin Resource Sharing"

   dc: tbl, you objected to "access control" so I took a follow up

   tbl: that's much better

   <DanC> close ACTION-203

   <trackbot> ACTION-203 Discuss Access Control misnomer with
   Interaction Domain staff closed

   <masinter> [31]


   <masinter> was just published yesterday

   <masinter> also


   <DanC> action-203: wg discussion of renaming spec includes
   68.html from anne


   <trackbot> ACTION-203 Discuss Access Control misnomer with
   Interaction Domain staff notes added

   <masinter> also




   skw: I've been chatting with them about de-referencing resources
   from widgets
   ... no need to make references into widget packages.

   <DanC> (sigh... I completely missed stuart's message, as it had an
   irrelevant subject of Re: Sketch of an idea to address
   widget/package addressing with fragID syntax and media-type defn )

   <Stuart> FWIW, we have removed all requirements for Widgets to be

   <Stuart> embeddable into Web documents. Widgets, as currently being
   defined by the

   <Stuart> Widgets 1.0 Family of Specifications, are solely packaged

   <Stuart> client-side applications that are authored using Web

   <masinter> there's a mailing list on this topic at
   n/ . I'd had an action item which I'm late on, to announce this
   mailing list in IETF


   <noah_waltham> Noah wants to remind himself that Jonathan sent an
   email on 22 Jan noting one more typo in Self-describing Web. It will
   be fixed before publication. Thank you Jonathan!

   <Zakim> timbl, you wanted to argue for allowing N of these and to

   masinter: I proposed getting a bunch of groups together that are
   separately working on this area and seeing if there is a common

   skw: the mailing list is a venue for that discussion

   <DanC> (I'm surprised Marcos agreed to talk requirements with other
   groups; it's already clear to me that ebook and widgets don't share

   <DanC> (for reference:
   [37] )


   <DanC> (I think a telcon might help a lot; these communities have
   disconnected vocabularies)

   <johnk> dorchard: do you have a link to the blog entry?

   <jrees> it's linked from my wiki page... i'll get that link

   I suggest getting a wiki together and even a comparison

   <johnk> thanks

   <jrees> [38]


   I note that Eran's wor of doing a blog entry comparing various
   discovery techniques was very useful.

   <jrees> (or belowmentioned)

   <DanC> ISSUE-20: Postel's law, dealing with invalid content,
   TInkerbell effect Larry Masinter


   <trackbot> ISSUE-20 What should specifications say about error
   handling? notes added

   Eran's blog post is


   <Norm> timbl, after the call, can you hang about for a few minutes

   meeting adjourned

   <Norm> 2:30 meeting cancelled. I can finally get off the phone. Tea.
   I want tea.

   <DanC> ah... tea... yes...

   tim, still some questions for you about your visit...

Summary of Action Items

   [NEW] ACTION: Noah to publish SDW finding, with 4 changes as noted
   in minutes of 22 Jan 2009 tag telcon due 29 jan 2009 [recorded in

   [End of minutes]

    Minutes formatted by David Booth's [42]scribe.perl version 1.134
    ([43]CVS log)


Received on Wednesday, 18 February 2009 02:52:40 UTC