W3C | TAG | Previous: 23 Jun teleconference | Next: 7 July 2003 teleconf

Minutes of 30 June 2003 TAG teleconference

Nearby: IRC |Teleconference details issues list www-tag archive

1. Administrative

  1. Roll call: SW (Chair), TBL, NW, DC, TB, CL, IJ. Regrets: RF, PC, DO
  2. Accepted minutes of 23 Jun teleconference
  3. Accepted this agenda
  4. Next meeting: 7 July
  5. The TAG does not expect to meet 18 Aug, 25 Aug, 1 Sep.

Next meeting with Voice WG?

SW: IJ and I met with some reps from Voice WG last week. They are revising some proposed text. They will circulate to IJ and me for review. If all goes well, I'd like to schedule some time with them to confirm it.
Proposed: Voice WG expected to join our 7 July teleconf for a small piece.
SW: If we have material from them we'll try to include them in next week's call.

2. Technical

2.1 Architecture document

27 June 2003 Working Draft of Arch Doc published.

  1. Action RF 2003/06/02: Rewrite section 5. Section 5 is expected to be short.
  2. Completed action DO 2003/06/02: Write up a couple of paragraphs on extensibility for section 4 (Done).
  3. Action IJ 2003/06/16: Attempt to incorporate relevant bits of "Conversations and State" into section to be produced by RF.
  4. Action PC 2003/06/16: Send second draft of AC announcement regarding TAG's last call expectations/thoughts and relation to AC meeting feedback.

    SW: I have no update on that action,.

[Ian]

IJ: New draft published last Friday. Comments are coming in on previous draft; I haven't read them. TB and DC did editorial pass.
DC: Balance between story and formal spec to my liking now. I'd like to add an illustration for the travel scenario.
TB: I have discomfort on the section on authority (section 2.3). I don't know why we have a section if not for programmers' benefits.
DC: I think this will be connected to an issue TBL is about to raise.
SW: We are likely to be looking at this document at ftf meeting.
[TBray]
For the record: I am substantially uncomfortable with http://www.w3.org/TR/webarch/#URI-authority because I don't understand what normative effect it would have on the behavior of implementors. If none, lose it. If some, specifiy it.
[DanCon]
keep in mind web arch impacts folks that read and write documents, not just coders, tim bray
[DanCon]
(not to say I'm 100% happy with #URI-authority section as written)

The following discussion took place at the end of the meeting.

[Ian]

TB: Things that we need to worry about (e.g., Chap 4 still missing)
[DanCon]
"2.3. URI Authority"
[Ian]
TB: I think we need time at ftf to talk about sections 2.3 and 3.2.1
[DanCon]
"3.2.1. Desirable Characteristics of Format Specifications"
[Chris]
3.2.2.2. Final-form v. Reusable conflicts in some ways with cp26
[Ian]
Check out:
http://www.w3.org/People/Bos/DesignGuide/introduction
[timbl]
Ok, so we have a commitment to put it on but nothing to reference yet.
[Ian]
IJ: I'd prune this section. Also, some of this text not specific to Web arch.
[Zakim]
DanCon, you wanted to note xlinkScope-23 has a home in "3.2.4. Embedding Hyperlinks in Representations"
[Ian]
DC: I see xlinkscope has a home in 3.2.4
IJ: I think CL is working on too much stuff right now.
CL actions include: error handling, content/presentation

2.2 Findings

See also: findings.

2.2.1 contentTypeOverride-24

[Ian]
IJ: Next steps? Does anyone want to read before we say "We think we're done"?
CL: Has the SMIL IG been contacted?
IJ: No.
[TBray]
Scenario 2 in Section 2 has funny formatting; grey surround-box misshapen
[Ian]
Action CL, NW: Read this draft by next week.
[Chris]
i will review it (skimmed but not read in detail)
[Ian]
DC: Should this go to public-tag-announce?
SW: I hesitate. People reading minutes will see this discussion.
[Chris]
if people want to discuss it that should happen on www-tag
[Ian]
Action IJ: Announce on www-tag that we expect to approve this finding in a week or so. Last chance for comments.
[Chris]
this is also relevant to the error handling issue

2.2.2 xmlIdSemantics-32

[Ian]
CL: I haven't completely updated. But nearly done. We should update with latest info.
SW: Should we offer an opinion?
[DanCon]
"No conclusion is presented." -- http://www.w3.org/2001/tag/doc/xmlIDsemantics-32.html
[Ian]
CL: The XML Core WG has been discussing this. I don't think we should pick a favorite from the TAG.
NW: I agree with CL on that point The Core WG is working on this.
IJ: Next steps?
Action CL: Revise this draft finding with new input from reviewers. 7 July Deadline.

The TAG did not discuss the following at this meeting:

2.3 Issues the TAG intends to discuss

The TAG expects to do a walk-through of the open and pending issues in order to determine:

  1. Is this issue nearly closed?
  2. Do we want to close this issue before going to last call?
  3. Do we want to discuss this issue at the face-to-face meeting?

See summary from Stuart.

[Ian]

TB: Pending, since RFC2396bis not finished. There was never a formal expression from TAG on those drafts. But every issue that arose we hammered out.
[Question of whether we should have a finding to close off the issue]
TB: I don't think we should.
CL: Mark your drafts as obsoleted.
[Chris]
http://www.textuality.com/tag/uri-comp-4
[Ian]
TBL: DC made a comment in a meeting with which I agreed - there are some axioms about resolution of relative URis that are not written in rfc2396bis.
DC: I do worry about that.
TBL: Normalization of "../" and "./" for example. Need a statement about invariants.
SW: I suggest you raise an issue with RF on the URI list.
TB: W.r.t. last call, I think we have a dependence on RFC2396bis. We are stuck with a reference to a moving target for now...
[Some agreement that not much need for ftf time on this issue.]
[DanCon]
15 | Yes | No # my summary
If there's spare time, I'd like to, but don't squeeze something else off.

[Ian]

DC: I think that LM did the comparison that we asked RF to do.
[TBray]
http://lists.w3.org/Archives/Public/www-tag/2003Feb/0208.html
[Ian]
DC: I think that msg merits discussion. Not sure whether in the path to last call. The business about why not create a new URI scheme is relevant here. Suppose ldap were being designed today. They could design a new protocol and make a new URI scheme. Or they could use HTTP as a substrate. The principle about don't make up new URI schemes and HTTP as substrate are related.
TB: While that's fair, I think that our comments are sufficiently general so that we don't need to change anything. If we want to provide information about when it is worth the cost, that might be ok.
[DanCon] I'd like to know Orchard's sense of propority of HTTPSubstrate-16

16 - Resolve for last call: No. Discussion at ftf: Spare time.

[Ian]
See Notes from Chris Lilley.
CL: "Ignorability" is something I'd like to discuss. If you get a file and it has an attribute in a namespace that you are supposed to understand, then that's an error. But if you add your own attribute in your own namespace, considered good way to extend. I think we should stay clear of extensions to XML.
TB: Some errors depend on application... In 3.2.1 of latest arch doc, bullet on attention to error handling.
[Chris]
dan - yes, the notes say that and give examples of harm from silent recovery and attempted recovery
[Ian]
TB: We might put something in section on XML...I would kind of be inclined to declare victory based on what's in 3.2.1
[Zakim]
DanCon, you wanted to say I'd like "silent recovery from errors considered harmful" in this last-call draft
[Ian]
DC: What I want in the last call draft is that "Silent recovery from error is harmful" to be in a box; critical for last call.
CL: Notes that I sent in gave some examples of bad consequences of silent recovery.
[CL cites example of browsers that consider </p> an error and treat it as <p>, so extra vertical space]
[Zakim]
timbl, you wanted to suggest we have said a bit too much about errors - one cannot tell people what to do if they do have an error.
[Chris]
aha - be careful what specs say about errors, that sort of thing?
[Ian]
TBL: I'm concerned about going too far in direction of saying how to design an application.
[Chris]
carefully distinguish from errors (fatal) and warnings
[DanCon]
I share timbl's concern. I still stand by "silent recovery from errors considered harmful"
[Ian]
[TBL cites example of inconsistent RDF; application-dependent scenarios]
[Chris]
at user option is no use in a batch job - good point TimBL
[Ian]
TBL: I don't like the SGML attitude of specifying the behavior of an agent. Just say what the tags mean. Don't tie down specs with overly narrow error-handling requirements.
[Zakim]
TBray, you wanted to agree with Dan about getting "silent failure considered harmful" into webarch before last call
[Ian]
TB: I think we have consensus that "silent recovery from errors" is probably bad behavior in the context of web arch.
[Chris]
have separate conformance reuirements for correct docs, correct generators, and correct readers
[Ian]
TB: I'd like to spend some time at ftf meeting on this.
[Chris]
and correct user agents a s asubset of readers
[Ian]
TB: XML's "halt and catch fire" might have been too much...
DC: I second talking about ftf.
CL: Not sure this is in the way of last call. Yes to discussion at ftf
[DanCon]
I think we might end up splitting it in half and closing one half.
[Ian]
TB: This one might not require a finding.
[Chris]
http://lists.w3.org/Archives/Public/www-tag/2002May/0124
[Ian]
CL: Specs, in the conformance section should be clear about when they are talking about documents, generators, and consumers.
[DanCon]
"What should specifications say about error handling?" http://www.w3.org/2001/tag/ilist#errorHandling-20
[Ian]
20: Schedule at ftf, try to close before last call.

[Ian]

SW: Last action was to write to HTCG and XML CG. I've had no feedback from either group.
CL: The XML CG has discussed. A task force to be created. The HTCG has discussed briefly. Some people seem interested....3/4 of a task force formed...
[TBray]
Suggest not on critical path for last cal
[Ian]
CL: Moving forward, but not much momentum.
Action CL: Ping the chairs of those groups asking for an update on xlinkScope-23.
SW: I set expectations that TAG would have a last look.
DC to TBL: Is what's going on with xlinkScope-23 consistent with your expectations?
[Stuart]
From: http://lists.w3.org/Archives/Member/tag/2003Mar/0104 " We believe that since we last considered this issue, there
has been substantially more input to the discussion, and thus we will commit
to taking up the issue once again and, should we achieve consensus, publish
that position as our contribution to work in this area."
[Ian]
TBL: I have the feeling that the way this will be resolved "nicely" is a new version of xlink or hlink that is simpler.
DC: My opinion is "no" and "no" (for 23)
CL: I agree with "no" and "no"

[Ian]

SW: I think we'll have this resolved for last call. Probably don't need to discuss at ftf.
[DanCon] on 24, I suggest yes for lc, no for ftf. (what Stuart just said)

[Ian]

CL: I was working on this one today.
[DanCon]
on 26, I guess I'm no for lc, yes for ftf
[Ian]
CL: I'd like to have some discussion before last call. And discussion at ftf since not yet discussed.
[DanCon]
(don't mind trying for 26 for lc)
[Ian]
CL: The finding I'm writing is a bit wordy.... If we all agree, could be slipped in; but don't think it needs to be in before last call. But I'd prefer.
DC: Worth a try.

[Ian]

TB: I think that after back and forth, we decided that the IRI draft was not cooked enough yet. I don't think we need to solve before last call. I don't think we need to discuss at ftf either.
[Zakim]
Chris, you wanted to talk about a new and related issue
[Ian]
CL: New and related issue - When do you use URIs for labels for things?
[Or should you use strings]
CL: I've started a writeup on this one...
SW: I hear "no" and "no" for 27.
TBL: IRIs extend 15 into IRIs. I think we could even work on this independent of IRI spec. Is this urgent?
[Chris]
yes its urgent, according to the XML activity
[Ian]
DC: I'd like ftf time on this one.

[Ian]

[No actions]
DC: Please add this to the pile containing 6, 37, 38
[TBL: And soon-to-be 39]
DC: "no" and "yes"

[Ian]

CL: I'd like to do a survey for this issue.
[DanCon]
why does this have a "resultion summary" if it's still open? I don't see how "draft" would resolve the apprent contradition between an issue being in "assigned" state and having a "resolution summary". not urgent.
[Ian]
[Dan, it should say "draft"]
Summary from CL:
http://lists.w3.org/Archives/Public/www-tag/2003Feb/0224.html
TBL: I'd like to add OGC to the entry for this issues list.
SW: No and No

[Ian]

SW: I hope to put out for TAG review this week.
IJ: Seem slike 31 is low-hanging fruit.
SW: No, Yes.

[Ian]

CL: I suggest we leave in pending. "No", "No"

[Ian]

No, no.
[timbl]
no no
[Ian]
CL: I'm happy to have discussion at ftf and write that up.
TBL: Connects to composable things.

TBL: No, no

DC: No, Yes. There is movement on this; I'd like some ftf time.

[Ian]

DC, TB: I'd like some ftf time on this.
[Norm]
+1
[Ian]
TB: I don't think impacts arch doc.
DC: Agreed
[Chris] no,no for me

DC: Cluster 6, 28, 37, 38

2.4 New issues?

The TAG does not expect to discuss new issues at this meeting.

  1. Summary from Stuart

2.5 Issues that have associated action items

3. Other actions


Ian Jacobs for Stuart Williams and TimBL
Last modified: $Date: 2003/07/07 19:07:55 $