W3C | TAG | Previous: 30 Jun teleconference | Next: 14 July 2003 teleconf

Minutes of 7 July 2003 TAG teleconference

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

1. Administrative

  1. Roll call: SW (Chair), DC, DO, NW, TB, IJ (Scribe). For the discussion about contentTypeOverride-24, Jerry Carter from the Voice WG attended. Regrets: TBL, RF, PC. Missing: CL
  2. Tentatively accepted minutes of 30 Jun teleconference with changes requested by Chris Lilley. Awaiting confirmation from Chris Lilley.
  3. Accepted this agenda
  4. Next meeting: 14 July. Likely regrets: NW.
  5. Face-to-face meeting agenda

    SW: I count on having an agenda for the TAG by next week.

2. Technical

  1. Review of text from Voice Browser Working Group and finding (contentTypeOverride-24)
  2. Review of comments on finding for whenToUseGet-7
  3. TAG comments on SW draft of finding for issue metadataInURI-31

2.1 Review of text from Voice Browser Working Group (contentTypeOverride-24)


JC: Thanks to SW and IJ for discussions last week and for comments on text. One point in contention was use of "type" attribute overriding server headers. We've concluded that that was inappropriate and the new language makes the headers authoritative. We are taking a direction that the HTML WG seems to be going with XHTML 2.0: using "type" as a kind of preference (accept header). Once representation returned, if no header sent, then type serves as a strong hint. Most of SW/IJ comments have been incorporated. There are a few remaining minor issues open.
TBray: Sounds like the results are clean and sensible. We should all be happy.
[JC gives quick overview]
TBray: Looks fine to me.
DC: Please clarify the semantics of the table.
JC: First three lines provide context; fourth line provides desired behavior.
DC: Declared media type, if I understand, is some combination of server type and preferred type.
IJ: I had initially said that dual usage of "type" was unnecessary; just use it as a preference.
I see potential editorial quibbles, but the thrust is architecturally sound, are we micro-managing?
IJ: But SW made a good point that it's cheaper to verify something's type than it is to determine the type of something.
DC: I"m not comfortable with "alleged media type". I'd be most comfortable with a test case. You are headed for PR?
JC: Yes, this is the only outstanding issue.
DC: Do you have test cases for this behavior?
JC: I think so. I would need to look at details. We have about 200 tests. I believe there's a test for this; I"d have to be sure.
DC: What would make me feel most comfortable is a test that shows a server returning text/plain and an implementation noticing the error.
JC: That's a reasonable request. There are probably 2-3 tests one would run for this (i.e., each of the special cases).
DC: It's not critical that the TAG endorse the final text. We were asked our opinion, we gave it, they tweaked their spec, and everybody's happy.
JC: Outside of this discussion, I owe some personal feedback regarding test cases. But I think the issues between the Voice WG and the TAG are resolved.
SW: I'd like to record our thanks to the Voice WG!
TB, DC: Absolutely.
The TAG wishes the Voice WG good luck!

[JC leaves]

More discussion on reviews of Client handling of MIME headers

  1. Action CL, NW 2003/06/30: Read the draft finding by 7 July. NW Done
  2. Completed action IJ 2003/06/30: Announce on www-tag that we expect to approve this finding in a week or so. Last chance for comments (Done)
  3. Completed action IJ 2003/06/30: Ping SMIL IG about this finding, which refers to architectural error in SMIL 2.0 (Done).
  4. See original summary of comments
  5. Comments from Norm Walsh
  6. See comments from Philipp Hoschka and comments from Rob Lanphier
  7. See suggested editorial changes from TBL
  8. Comments from Stuart Williams


Comments from Norm Walsh

NW: My comments were mostly editorial nits; fine on the whole.
NW: In section 6, bullet 3, I would support "MUST NOT": "Specifications MAY include hints about server headers but SHOULD NOT include requirements that a client override these headers without involving the user."
DC: Seems strange to constrain the spec. If you don't have the user's consent, you can't hold them responsible. I don't think this is the best advice we can give to spec writers .If a user agent wants to do something different than the published protocol, they have to do so with the consent of the user, otherwise they are not acting on behalf of the user.
TBray: Any time somebody writes this sort of thing in a spec, they can be sure they'll get static from the TAG.
DC: Will we read every spec? We can just say "The server mime type is authoritative."
TBray: Then we should say nothing about what specs do..."SHOULD" is misleading.
NW, DC: Change to MUST NOT in bullet 2.
IJ: I can live without bullet 3 since section 5 covers this.

Resolved: In section 6 of finding, drop bullet 3 and to change SHOULD NOT to MUST NOT in item 2.

Comments from Rob Lanphier

TBray: I think some of RL's comments about servers are useful, e.g., default content type is harmful.
IJ: Seems like a reasonable addition to talk about how to address this issue in practice.
DC: Don't talk about "harmful", talk about "risks of default content type" The right answer in case of misconfigured server is not to sweep it under the rug.

Suggested editorial changes from TBL

TBray: Unless TBL can provide some examples of the impact on programmers, what's the utility of this?
DC: To me this look editorial and is not a huge improvement. Putting "the" in your face doesn't help things. I'd prefer that IJ return to TBL and ask for more motivation.
TBray: I like TBL's example of showing how even self-describing content not sufficient (or may provide wrong info) about author's intent in some cases.
Action IJ: Produce a new draft of the finding on MIME headers that takes into account comments from Rob Lanphier, TBL, NW, SW (editorial), and above resolution.

Comments from Stuart Williams

SW: I'd prefer "state of the resource" over "meaning of the resource". I think that "meaning of the resource" is more complex than its state. Also, the second security example didn't seem so credible to me.
DC: "State" of the resource is ok.
IJ: I'll produce a new draft for next week.
Ian: I think asked you to send a note to the SMIL folks saying "thanks, we're taking your input seriously & will redraft"
Action IJ: Follow up with SYMM IG saying we'll take into account input and produce new draft of this finding.

2.2 Review of comments on whenToUseGet-7 finding

Summary of actions:

  1. Action DO: Send text that DO and Noah M. can agree to to tag@w3.org.
  2. Action IJ: Produce new draft of finding with (1) change based on LM comment, (2) text from DO/NM (3) don't address PUT in this finding, but add comment about new issue putMediaType-38 (second point in summary of comments).

2.3 TAG comments on SW draft of finding for issue metadataInURI-31


(hmm... no story atop 4 July 2003 draft of finding)
DC: I think the finding may be getting too long. I was hoping for three paras that say "Don't peek into URIs."
"People and software making use of URIs assigned outside of their own authority (i.e. observers) MUST NOT attempt to infer properties of the referenced resource except as licensed by relevant normative specifications or by URI assignment policies published by the relevant URI assignment authority."
SW: Can one make use of information published by W3C (in a policy)?
DC: Shorten 3 to "Don't peek inside the URI." It's always safe to not peek inside the URI. On Scheme component designators and wsdl components....
the W3C tech reports naming policy isn't based on the URI alone. there's no guarantee that http://www.w3.org/TR/2003/WD-snort-20040707 is a W3C working draft. If you know that it is a WD somehow, you can learn stuff from the URI. but you can never just peek into the URI alone.
DO: DC's point (don't peek in URIs) is in conflict with what some groups are trying to do.
SW: Does this draft need more work before sending to www-tag?
DO: I agree with DC that it seems a little long.
NW: If I understand DC correctly, sounds like DC is saying that if I know XQuery is a Working Draft, then I can find meaning in WD in URI. If I saw a URI ending in "WD-".... could I infer that it's a WD?
DC: No You have to look at the TR page as well. Once you have found something on the TR page, there's quite a bit you can go on. If you start from the TR page and find the URI, that's fine. But if you find the URI on the floor somewhere, you can't go on that alone.
That http://www.w3.org/TR/wd-foo exists does not imply foo is a WD unless and until it appears *as a link* on the TR page.
DC: Just don't peek in the URI.
DO: I don't want the finding written that way.
DC: A story would be useful. If you want to treat the entire topic, then start with a story. It's not clear why someone wants to read all this stuff. I don't think IJ is hot on the trail of a story. I don't think this comes up very often...may not deserve a real story.
SW: I will take a crack at writing something short before floating on www-tag.
NW: Perhaps DO could write a couple of paragraphs to motivate why one would want to do this.
which wsdl group posting?
DO: I'll write up something in response to SW's draft.
DC: I'm not sure putting burden on SW is appropriate. Maybe just send what you've got and say there isn't agreement. Promote third bullet to abstract.
DO: Amazon publishes policy for query parameters.
DC: Perhaps cheapest approach is to open up to community.
treating the whole forms interface convention is more than I would have expected for a finding on issue metadataInURI-31
DO: Where do others stand on this?
NW: I'm definitely on the fence. I think there's a lot of merit in "don't look inside the URI", but I understand what the WSDL WG wants to do.
DC: Don't put version numbers in URIs was the original question.
TBray: I think the answer is still "probably not"
Action SW: Create new draft based on input today and send to www-tag.
Action DO: Send rationale about why WSDL WG wants to peek inside the URI.

2.4 Findings the TAG did not discuss

See also: findings.

Next steps on these findings:

2.3 Architecture document

27 June 2003 Working Draft of Arch Doc published.

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

About Paul's second draft of comments to AC.

the question of who sends it to ac-forum is a good one
Issue is: who sends this thing, and to whom
SW: want's TimBL's sign-off. I would want to sign it from both of us.
DC: Don't put TBL's signature at bottom of something he hasn't read.
TBray: I propose that SW find TBL and get him to sign off.
Action SW: Try to get TimBL to sign off on Paul's text. If SW able to reach TBL, SW/TBL send to AC as co-chairs. If not, have IJ send on behalf of TAG.

2.4 Issues the TAG did not discuss

2.5 Issues with action items

2.6 Issues the TAG intends to discuss at face-to-face meeting


Qnames, fragments, and media types

RDDL, namespace documents

2.7 New issues?

The TAG expects to discuss these on 14 July.

  1. Visibility of Web Services, raised by Mark Baker
  2. Character model conformance, raised by TBL
  3. Meaning of URIs in RDF documents, raised by TBL

3. Other actions

Ian Jacobs for Stuart Williams and TimBL
Last modified: $Date: 2003/07/07 23:33:41 $