W3C | TAG | Previous: 9 Feb teleconf | Next: 2 Mar

Minutes of 23 February 2004 TAG Teleconference

Nearby: IRC | Teleconference details · issues list (handling new issueswww-tag archive

1. Administrative (20min)

  1. Roll call: DO, MJ, TBL, RF, CL, SW, IJ, TB (Scribe), NW, DC, PC
  2. Accepted minutes of the 9 Feb teleconf
  3. Accepted this agenda?
  4. Proposed next meeting: 2 Mar 2004
  5. Resolved to cancel 8 Mar 2004 teleconference (and reconvene 15 Mar)
  6. Finalize meeting dates in August in Ottawa?

    Action PC 2004/02/09: Propose August ftf meeting dates. (done)

    The TAG tentatively plans to meet in Ottawa 9-11 August.

1.1 Technical Plenary

  1. Liaisons update

1.1.1 Plenary day planning

[TBray]

SW: will open with 10 minutes, DC will do namespace docs...
DC: some risk for DC
SW: idea is presentation-light, discussion-heavy
[DanC_desk]
(TBray, maybe I'll call you to think a little bit about this namespace documents discussion)
[TBray]
SW: DO to motivate extensibility/versioning
DO: short pres showing where we're at; motivate discussion by asking hard question; e.g. should Schema 1.1 fix this prob, should we introduce "must-understand" somewhere
[DanC_desk]
sounds good, DO
[TBray]
DO: will write up in email form, send to group to see if appropriately motivating
SW: also, web identifiers
TB: not enough time for three discussions
SW: can RF motivate identifiers?
RF: yes, but didn't understand hints. Given schedule... i.e. an indication of what I should prepare
DO: suggests complex identifiers for multi-processing-step docs
TB: have 2 discussions not 3
Drop namespace docs?
TB: boring
DC: audience still interested
SW: Mechanics? queue handling? I will aim to moderate, pick questions from floor & panel. done for now?
Other notes:
  1. Intro to TAG and Web Arch (Stuart)
  2. Namespace documents (Dan C)
  3. Namespace documents (Dan C)
  4. Web Identifiers (Roy F? Qnames, Ids, FragID, URI andIRI)

    Action SW 2004/02/09: Find a volunteer to discuss identifiers at Tech Plenary

  5. Boilerplate:
    1. Outline issue
    2. Why TAG interested. What's difficult?
    3. Known positions around issue
    4. Encourage discussion/debate

1.1.2 TAG ftf meeting agenda

  1. TAG 2 Mar 2004 ftf meeting page

CL: Request triage of action items to close.

NW: Ian, I believe I've finished three of my four actions (all except asking the schema) and I should have a fifth/second: responding to Hammond's last call comment.

2. Technical

See also open actions by owner and open issues.

2.1 Update on findings

2.1.1 qnameAsId-18

[TBray]

SW: Pending input from TB & TBL
CL: I looked at it again, didn't see problems
TBL: don't want to hold it up
TB: likewise
... discussion trying to re-establish context...
[Ian]
http://www.w3.org/2004/02/09-tag-summary.html
[See comments at that meeting from TBL]
[DanC_desk]
"[timbl] On Qname finding: I think NW should make more of the algorithm that one uses to determine the binding when looking at elems and attributes."
[timbl]
"There is no single, accepted way to convert QNames into {URI, local-name} pairs or vice versa." (para n-3 in 4.1)
[TBray]
CL: Revisiting this point...
[Chris]
xptr fails to inherit scope into the xpointer
[TBray]
NW: if you use a prefix in an xpointer and there is no xmlns thingie in the xpointer, you can't use the in-scope prefixes fo the xml doc
[Zakim]
Chris, you wanted to offer to be atthe wsd-wg meeting if needed and to raise a non-holdup technical point
[TBray]
TB: seems like NW has a factual statement of what things are like. are we inclined to criticize it?
DC: seems that what XPointer did is right
TBL: XPointer is a mini-language can be used anywhere... nothing to do with XML in principle
[Chris]
if they had done it differently, you would need to rewrite an xpointer to move it to a different context
[DanC_desk]
quite, chris
[TBray]
TBL: problem would be if in another XML vocab e.g. WSDL they had another way of mapping... e.g. by looking in a schema; there is an extra cost .. if you take Norm's statement literally this suggests that there shouldn't be a single way; lower cost if everyone uses the same one
NW: Don't think I said that. Should I modify to say that this is commonly used, and should be used
[Chris]
I see we are now getting discussion on this doc, brought on by the imminent decision to approve the finding
[Norm]
lol
[DanC_desk]
why is that remarkable, chris?
[timbl]
It is not as though we have a solution here.
[TBray]
TB: in the context of XML docs, it would be nuts to invent a new way of doing it, so the finding should say that in particular it's OK for XPointer because it exists outside the XML context...and the finding could usefully call that out too. e.g. take XPointer as an example
CL: This came up in the SVG development. We had an attribute that pointed to another to animate it; could be in a different subtree/doc. Question is: in which context do we do this, the pointer's context or the target's context
[Zakim]
Chris, you wanted to relate another way to resolve them (that was rejected, and rightly)
[TBray]
CL: and in this case, the right thing to do was in the context where the prefix actually was
[Zakim]
mario, you wanted to say: isn't the ns non-inheritance thing of XPtr conflicting with the inheritance mechanism present in XPath
[TBray]
MJ: when I embed in an XPath... the xpath inherits but an xpointer doesn't
NW: but then XPath expressions aren't URIs
PC: and XPath doesn't define that, XSLT does
[Chris]
reminds myself to check what dom3 xpath does by way of a host language
[TBray]
NW: XPath context has to be init'ed, XSLT & XQuery use mappings in containing doc
[Chris]
it might be useful to capture some of these cases into the finding. If its unclear to us, its unclear to others too.
[TBray]
TBL: might be useful to look at some of these cases and discuss the cost. Might be able to rebind prefixes.....
[Zakim]
timbl, you wanted to quantify the cost: The ability (or not) to be able to re-generate prefixes, to be able to determine two document fragments to be the same, without application knowledge, but with schema knowledge.
[TBray]
NW: no; can't rebind prefixes in the general case
[DanC_desk]
(I gave up hope on rebinding prefixes long ago)
[TBray]
TBL: can in some cases
NW: that special cases isn't large enough to call out. Will amend the wording from sim ple statement of fact to a recommendation to do the right thing will have that done before Sat
[timbl]
(I rebind prefixes all the time in RDF of course)
[Chris]
+1 for Norm's suggestion
[TBray]
TB: wordy +1
ACTION Norm to rewrite simple statement
[DanC_desk]

thanks, again, norm.

2.1.2 contentTypeOverride-24

[Roy]
I have read "Authoritative Metadata" draft 18 Feb 2004 and all of my concerns have been addressed. The document looks great. Thanks Ian!
[TBray]
see http://www.w3.org/2001/tag/doc/mime-respect-20040218.html
SW: Anyone else seen it?
DC, TB: prepared to defer to Roy
[Chris]
looks good to me
[Norm]
yes
[TBray]
SW: Resolve to accept?
RESOLVED to accept that finding
IJ: now I publish it, right? I would like to make some mods to webarch based on the good work done here
[Chris]
4.2 Self-describing data and Risk of Inconsistency is very good, glad to see that.
[DanC_desk]
hmm... arch doc changes? such as?

IJ: E.g., clarification that sender of metadata not always server.

2.2 Follow-up on namespaceDocument-8

[TBray]

DC: I sent email to address comments
[DanC_desk]
(my last message on this issue http://lists.w3.org/Archives/Public/www-tag/2004Jan/0026.html)
[TBray]
TB: also unconvinced by pleas for XLink version
DC: asked for pointers to use cases for XLink version, thinks people sent them
[DanC_desk]
yup, 0026 points to my XSLT/RDDL work
[TBray]
TB: missing PC's action item. No interest in a normative RDDL note?
PC: Wrong. AC asked us to turn it into a note and/or give it normative status
CL: and we decided that RDDL wasn't always appropriate
DC: but that doesn't change the status of RDDL, which we officially think is useful
[DanC_desk]
(not officially yet, actually)
[TBray]
PC: and the direction from the AC was that they wanted something to point to
[DanC_desk]
(ah yes... bristol decision makes it official, I suppose)
[Zakim]
timbl, you wanted to mention push-back agiuanst TAG doing rec track spec here.
[TBray]
TBL: within the team, there's been pushback against TAG doing rec-track work. Members can't join in at will
[Zakim]
TBray, you wanted to volunteer what work I'm prepared to do
[TBray]
TB: is prepared to go on editing RDDL as long as it stays reasonably simple
PC: discomfort with team-only discussions; would like team concerns reported so he can address them, e.g. by pointing to our charter where it says we can do rec-track work
SW: we'd only really discussed taking this forward as a NOTE
[timbl]
Ok
[Chris]
suggest TimBray takes it if he is motivated
[TBray]
TB: we need the finding, Paul's not getting it done, does somebody else get it?
[timbl]
PC: To motivate the use of one or more formats.
[TBray]
DC: not interesting because webarch says enough
[DanC_desk]
(finding could talk about content negotiation, etc.)
[TBray]
PC: finding would provide more motivation and discuss alternatives
[DanC_desk]
I don't think my GRDDL background document has much to offer on issue 8. it's more on issue 35
[TBray]
see http://www.w3.org/TR/webarch/#namespace-docs
PC: but if webarch has enough, maybe we don't need finding?
TBL: could build DC's GRDDL stuff into a finding?
DC: yes it's a finding, but on issue 35 not 8
[timbl]
Ok, I accept that issues 35 and 8 have small overlap
[TBray]
PC: will write something for Friday so as to get on F2F agenda,
[timbl]
Maybe we should drop the finding, if the arch doc actually covers it.
[TBray]
PC: it'll be an outline-form thing, if people like it we'lll add the meat
TBL: this really worgh doing, not unnec work for PC?
PC, DC: seems worthwhile
[timbl]
I hope someone brings a video camera to the face-face, now TimBray and I will neither be there.
[TBray]
DC: people using XLink version, are we worried about that?
[Chris]
who was using it? eric van der vlist, as i recall
[TBray]
Just Eric as far as I know
[DanC_desk]
[[[
In the meantime, I think that http://examplotron.org,http://xsltunit.org and even http://rddl.org/ (version 1) are goodexamples.
]]
[TBray]
Henry Thomson also using it....
[Stuart]
Henry's message http://lists.w3.org/Archives/Public/www-tag/2004Jan/0044.html
[Ian]
TBray: There are some benefits to the xlink version, but also more complex.
[TBray]
TB: it's a cost/benefit trade-off
TBL: if we make RDDL official, would JBorden make some commitments. We want a persistence commitmment, along the lines of keep it commitment or sign over to W3C, or to a trust or some such.
TB: ACTION TBray to check with Jonathan about persistence policy if RDDL hosted there.
TB: anyone here want to go back to XLink?
CL: it had more deployment than we thought, and we can make RDF equally well from eiither
DC: TBray, do plan to change RDDL.org to use the attribute version?
TB: yes
TBL: will the namespace doc at RDDL.org enable fetching the RDF
TB: yes
DC: mime-type issue is nontrivial
CL: if you serve as xhtml+xml, should be OK in modern browser
[DanC_desk]
(anybody got a handy test page for application/xhtml+xml ?)
[TBray]
PC: serve as what?
CL: application/xhtml+xml
[DanC_desk]
http://www.w3.org/People/mimasa/test/xhtml/media-types/
[TBray]
DC, CL: it's OK to use foreign namespace attributes in an xhtml+xml doc
[Zakim]
TBray, you wanted to get unlost
[TBray]
TB: I favor the shorter attribute-based syntax
TBL: can mention previous syntax?
[Ian]
Summarizing actions:
Action PC: Give points for potential finding
Action TB: Continue working on draft and to get statement from Jonathan re: persistence at rddl.org
Action TB: Add pointer to previous syntax in the Note
[DanC_desk]
(we talked about the syntax of the http://rddl.org/ document changing too, but that can be asynchronous)
[TBray]
PC: do we now go back and change the language in webarch about alternate forms
CL: yes, maybe
DC: yes, but in some cases the Owl doc is optimal
SW: likes the Bristol compromise
[DanC_desk]
yes, what's in the arch doc is right
[TBray]

TBL: likes the status quo

2.3 Web Architecture Document Last Call

2.3.1 Peter Patel-Schneider

  1. Comments from Peter Patel-Schneider
  2. Many of the comments concern notions of authority and ownership; see related comment from David Booth about ns documents.

[TBray]

SW: Re: Strickler's comments
[Stuart]
Section 4.5.4: "It is disappointing to see the TAG continuing to promote the idea that any semantics associated with a URI used as a namespace name has any relation whatsoever to the semantics of terms grounded in that namespace."
[TBray]
DC: disagrees with the notion that nothing more need be said, because saying nothing has had negative consequences
TB: Patrick's position has for a long time been that namespaces are pure syntax/punctuation
PC: seems to disagree with our opinion that human-readable doc is useful
... discussion too fast for scribe...
is it the case that namespace is just another ersource
DC: I don't think we disagree with him, he's just saying that we don't want to discuss this at length. But we think discussing it at length is useful
[Ian]
DC: Patrick's point has technical merit; but there is social benefit to explanation.
[Stuart]
Defn of NS Document from Webarch: Namespace document
The resource identified by a namespace URI.
[Chris]
sounds like agreement to me
[DanC_desk]
hmm... indeed, the glossary entry is goofy
[Ian]
DC: Fix - it's what you get back when you dereference.
[TBray]
TB: strongly disagree with PS's paragraph beginning "Furthermore, because..."; I think that when you use a URI as a namespace name, you've created a resource and you better make sure it has something to do with the vocaublary
... discussion of the glossary entry....
TB: if you disagree with glossary entry, provide alternate text
[Chris]

the detailed reading is indeed gratifying

2.3.2 Sandro Hawke

  1. Comments from Sandro Hawke

2.3.3 Dominique Hazael-Massieux

  1. Comments from Dominique Hazael-Massieux
  2. Comments about metadata
  3. Comments about conformance

2.3.4 Danny Weitzner

  1. Comments from Danny Weitzner

2.3.5 David Booth

  1. Question about ns docs providing definitive info

2.3.6 Sergio Rodrigues

  1. Question about XDI and XRI w.r.t. Webarch

2.3.7 Jacek Kopecky

  1. AWWW LC comments

2.3.8 Martin Dürst

  1. principles, etc.

The TAG does not expect to discuss the topics below this line.

4. Follow-up on Internationalization Issues

5. Status report on these findings

See also TAG findings

6. Other action items


Ian Jacobs for Stuart Williams and TimBL
Last modified: $Date: 2004/02/26 15:41:07 $