W3C | TAG | Previous: 25 Nov teleconf | Next: 9 Dec 2002 teleconf

Minutes of 2 Dec 2002 TAG teleconference

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

1. Administrative

  1. Roll call: All present. SW (Chair), TBL, DC, TB, CL, PC, NW, DO, RF, IJ (Scribe).
  2. Accepted 18 Nov minutes
  3. Accepted 25 Nov minutes
  4. Accepted this agenda
  5. Accepted draft summary of TAG work from previous month (with changes suggested by DC and additions from this meeting). IJ to include TB's draft finding on URI comparison in summary.
  6. Next meeting: 9 Dec 2002

1.1 Completed actions

1.2 Meeting planning

1.3 Other business?

The TAG discussed its slide presentation from the W3C Advisory Committee meeting.

2. Technical (75min)

Possible new issues:

  1. SOAP and XML internal subset
  2. Binary XML
  3. Metadata in URIs
  4. Postponed issues
  5. Arch Doc/Findings

2.1 SOAP and XML internal subset

See message from Paul Grosso


yes, please; I don't want to take this up until the XMLP WG has responded to a "don't subset XML" request.
DO: I think this is an important arch issue. I think it should have been sent earlier to XMLP WG. But having said that, the topic of subsetting has come up before. I'm on the record of wanting a next generation of profiles. I think the TAG ought to bring up this issue and consider the arch ramifications of profiles of XML specs.
TB: I agree with DO. The IETF's BCP says "don't do this." For the case of SOAP, I think they have overwhelming technical arguments for their design choice (namely, avoid risk of denial of service attacks). I think that in general, subsetting XML is probably not wise for the reasons cited by the IETF authors. There is a recurring desire of some groups to do this; that signal should be looked at.
Timmit, you wanted to agree we should take up issue and to suggest that original WG should be heavily involved and/or in charge when a profile is made of any spec.
TBL: I agree we should accept the issue. While there is a WG that is responsible for this work, I think it's important that that WG do the work. We should not establish the precedent that one group can profile the work of another (notably cross-organizational boundaries). We can discuss it, with an option to return to the XML Core WG with a request to produce a profile.
DanCon, you wanted to express a preference for having PaulG/XMLCore make a request to XMLP WG before we accept this
DC: If we accept this as an issue, can we immediately contact both WGs to ensure that they know they are represented?: One possibility: do this by email or in a teleconf. I would prefer that Paul write to the XMLP WG and get their reply on record.
NW: There's a lot of editorial work, not much technical benefit, unclear political ramifications of such an exercise. By "political" I mean that it's not clear what buy-in would be obtained from vendors and parser authors, etc. I'm not sure that's really "political" but it's more than purely technical.
DO: I think that Paul Grosso should ask the XMLP WG for their rationale, and that the TAG is interested in that reply. I believe that Chair of XMLP WG is interested in providing information on this topic.
PC: On the IETF BCP - does this apply when XML used as the basis of a protocol?: TBL talks about profiles as though they were bad; but profiles happen all the time within W3C.
I wonder if there are at least two profiles: "non-protocol" and "protocol".
PC: I'm not sure that the TAG can do anything about on group profiling (or not profiling) the work of another group.
I don't think TimBL suggested lockstep; I think he just meant that if XMLP wants to profile XML, the WG working on XML should get the right of review
PC: There's a long history on this topic (going back to Sep 2001, at least, see message on xml-dist-app) regarding SOAP. I think it is appropriate to tell Paul G to talk to the XMLP WG. We can give him some pointers to the public record.
TB Proposal:
  1. We should officially respond to PaulG saying that there is some history and that it would be appropriate to direct his query to the XMLP WG to ensure that the evidence is brought out for review.
  2. Propose TAG adopts subsetXML issue, based on the fact that XML doesn't provide a means for subsetting. Some people (like me) think that it's bad to subset XML. But some groups still want to do this, and some groups have good reasons for doing so.
hmm... I thought the subsetting XPath case was directly relevant. the name "subsettingXML" seems exclusive of that.
Dan, what is "XML"?
a language defined in the XML 1.0 recommendation, I think.
Is XML=XML 1.* + namespaces + xpath + dom + xquery + xslt?
DaveO: XML is http://www.w3.org/TR/REC-xml
TB: The XMLP WG has evidence that subsetting will be sometimes necessary.: It's not reasonable for SOAP 1.2 to wait for a revised XML.
DaveO: XML++ might include namespaces, base, et. al.
DO: Friendly amendment to TB's (2). Not just an issue of subsetting XML, but rather among the family of XML specs.
TB: Retitle as profileXML.: Change wording "XML family of specifications"
DC: "Profiling W3C specs" would be fine.
You don't conform to XPath. You conform to XPointer or XSLT.
DC: Flavors of a language are evil. Sometimes you need profiles, but there is a cost to interoperability. Profiles are to be avoided.
RF: What you want with a profile of XML is to make it possible to implement software.
I agree with Tim's amended resolution of this item but I would like to see a clear statement of the "XML family of specification" issue.
RF: General purpose servers implement HTTP differently from specific-purpose servers. There are limits on URIs, size of request header. Apps need to be able to define these things on their own. Not limits on the protocol, but limits on the implementation of the protocol.
Re XPath, I guess you could also conform to the new DOM API + XPath.
TB: I agree with DC - one of the good things about XML historically is that it's much more option-free than other specs. Clearly this approach is running into trouble. I've put a stake in the ground about which specs to group together (XML, namespaces, base) so that profiling not necessary.
TBL: Things will break if some parsers understand entities and others don't.
hmmm... sounds reasonable
Timmit, you wanted to say that http example is more -- what happens if an http server doesn't implement HEAD?
issue profilePlussesAndMinuses-NNN
issue profilesNecessaryEvil-NNN
PC: Wasn't this on the XML Core WG agenda at some point?
DO: Yes, they are chartered to do this.
TB: Maybe it suffices to say to the XML Core WG that we think this should be moved up their list of priorities.
NW: No one I know of is chomping at the bit to address this; seems like a lot of work, without much promise of payoff. If we want this work done, we should ask the Core WG.
TB: Don't phrase this as "Do XML 2.0". If we think there's a problem here (and I think evidence suggests there is), we could profitably invest some time in how we get a solution. Will be hard to disentangle tech from process issues.
DO: This issue has also come up in WSA WG.
NW: The major issues here are not technical.: The Core WG has discussed this.
DO: What information can be conveyed here?
TB: Let's toss this out into www-tag.
NW: If we want to engage the Core WG, we should invite Paul Grosso to a meeting where this is discussed.
DanCon, you wanted to propose: profilesNecessaryEvil-NNN
SW: I have concerns about our communications with other groups.
DC: We should accept issue and PC/DO and NW should ask the groups how they want to be represented here.
IJ: Title of this issue?
I don't want us to send an appeal to www-tag on this front since I want the negotiatiation with the Chairs and WGs to occur first.
TB: Whither and how to profile W3C specifications in the XML Family
DC: I object to "in the XML Family"
DO: I feel strongly about "in the XML Family"
TBL: I feel that the profiling issue applies to other issues as well.
he didn't say "feel strongly"; he (DaveO) observed that the XML family is what we've been talking about
DO: I'd like to examine the issue w.r.t. the scope of things in the XML family of specs.
TBL: If the comment is "necessary evil" then it applies to all our specs.
CL: We have two issues (general and specific).
SW Proposed: Accept profilesNecessaryEvil-NNN as new issue.
DO: I object.
I don't like it either
Objections: PC, CL, TB.
TB Proposed: xmlProfilesNecessaryEvil.
CL: I don't like "necessary evil"; presupposes an outcome.
Proposed: xmlProfiles.
issue xmlProfiles-NNN: When, whether and how to profile the XML family of recommendation
DC: Abstain.
like daves
Resolved: xmlProfiles. DC Abstains.
Action IJ: Add to issues list xmlProfiles-NNN. TB suggests title "When, whither and how to profile W3C specifications in the XML Family"
Action DO: Talk to XMLP WG about this new issue.
Action NW: Talk to XML Core WG about this new issue.

2.2 Binary XML

See messages from Robin Berjon, Paul Cotton (member only), Don Brutzman (member only)


For accepting: DO, DC.
Objections: TB
Abstain: NW, RF.
RF: I abstain, mostly because I wouldn't call it XML....
TB: Exactly, if it's binary, it's not XML.
its not the XML serialisation format, true
SW Proposed: Adopt binaryXML-NNN as an issue.
Objections: TB
Abstain: NW, RF, SW, PC
proposed - binaryXMLInfoset
PC: My rationale - I'm not sure what the community is asking for.
CL: Discussion started before I could send crisp problem statement.
I would like to be on the record as to why I would have supported this
Supports binaryXML-NNN: DO, TBL, DC, CL
I would like to take up this issue because it has been raised by so many parties too often, and no statment one way or the other exists about it. The community deserves such a thing.
Action CL: Write up problem statement about binary XML; send to www-tag.
TBL: Here's why I think we should take it up - it's been raised by a lot of people (e.g., Web3D, who are users). The XML community has ruled it out of scope. If the TAG's conclusion is that it's better to do Y than binary XML, then we should say so clearly. If the answer is so obvious, we should state it clearly. If it's not, then we should unearth it and deal with it.
considering changing my vote
DO: Also an issue in the Web Services community.: I think the TAG could help out in this area.
CL: For SVG we said "use gzip" but the mobile folks said that wasn't good enough; they have to store strings with whitespace preserved. They end up having 2 copies of the data.
(because the DOM allows access t the original strings)
TB: I am profoundly against the notion of binary XML in general. However, having listened here, it's apparent that it's an issue that won't go away.: If it's a bad idea, we should say why and tell people how to solve problems in the real world.
SW Proposed: binaryXML-NNN as a new issue.
Objections: None.
Abstain: RF, NW, PC
Support: DC, CL, TB, DO, TBL, SW
Resolved: Accept binaryXML-NNN as a new issue.
Action IJ: Add binaryXML-NNN to issues list.

2.3 Metadata in URIs

See message from Ossi Nykänen.


DO: I'm interested in the issue of versioning resources.: E.g., namespaces and versioning.
TB: Notion of encoding metadata in a URI is broken. Versioning has application-specific semantics.
Now RDF for metada is a good idea...
memories of VMS filenames with ;version in the filename
CL: If we universally thing this is a bad thing to do, we should say so loudly.
SW Proposed: Accept medataInURI-NNN?
IJ: Could be short if universal response is "no".
Resolved: Accept issue matadataInURI-NNN with note that TAG thinks the answer is "no" and will explain what to do instead.

2.4 Postponed issues

  1. Status of URIEquivalence-15, IRIEverywhere-27. Relation to Character Model of the Web (chapter 4)? See text from TimBL on URI canonicalization and email from Martin in particular. See more comments from Martin.
    1. Action MD 2002/11/18: Write up text about IRIEverywhere-27 for spec writers to include in their spec.
    2. Action CL 2002/11/18: Write up finding for IRIEverywhere-27 (from TB and TBL, a/b/c), to include MD's text.
    3. Action TB 2002/11/18: Write a finding for URIEquivalence-15 on IRI relation to URI, different levels of equivalence. Done
  2. namespaceDocument-8
    1. Action NW 2002/11/18: Take a stab at indicating pros and cons for the various RDDL/RDF/Xlink designs arising from TB's RDDL challenge.
  3. rdfmsQnameUriMapping-6
    1. The Schema WG is making progress; they will get back to us when they're done. See XML Schema thread on this topic.
    2. Action IJ: 2002/11/25: Update rdfmsQnameUriMapping-6 to indicate waiting on WSDL WG. Done.
  4. uriMediaType-9:
    1. Action DC 2002/08/30: Write a draft Internet Draft based on this finding (Deadline 2 Dec). This action probably subsumes the action on TBL to get a reply from the IETF on the TAG finding. Done, TAG only.

      Action DC: Point to this draft on tag: "A Registry of Assignments using Ubiquitous Technologies and Careful Policies." This action was corrected per 9 Dec 2002 TAG teleconf. The action was to point TAG to document, not www-tag.

    2. Resolved: The TAG thanks Mark Baker for his contributions to this draft!
  5. fragmentInXML-28 : Use of fragment identifiers in XML.
  6. xlinkScope-23 (5 minutes)
    1. Action SW 2002/11/18: Organize a special-interest teleconf for discussion of this issue on linking. Pending; see email from SW (TAG-only).

2.5 Findings in progress, architecture document

See also: findings.

  1. Findings in progress:
    1. deepLinking-25
      1. TB 2002/09/09: Revise "Deep Linking" in light of 9 Sep minutes. Status of finding?
    2. URI Comparison.
      1. Resolved: Link to TB's draft finding from findings page. Action IJ: Link to this from findings page.
  2. 7 Nov 2002 Arch Doc
    1. Action CL 2002/09/25: Redraft section 3 based on resolutions of 18 Nov 2002 ftf meeting.
    2. Action DC 2002/11/04: Review "Meaning" to see if there's any part of self-describing Web for the arch doc. Done.
    3. Complete review of TBs proposed principles CP9, CP10 and CP11

Ian Jacobs for Stuart Williams and TimBL
Last modified: $Date: 2002/12/09 22:02:54 $