W3C | TAG | Previous: 24 Jun | Next: 8 July

Minutes for 1 July 2002 TAG teleconference

Nearby: Teleconference details issues list www-tag archive

1. Administrative (15min)

  1. Confirm scribe: IJ
  2. Roll call: TBL, CL, SW, RF, IJ. Regrets: DO, PC, TB, NW, DC
  3. Accept this agenda
  4. Next meeting: 8 July. Regrets: TBL
  5. Accepted 24 June minutes
  6. Confirmed status of completed actions

1.2 Completed actions?

  1. IJ: Ask TBL to take ownership of issue xlinkScope-23. Done. Refer to TimBL proposal for when to use XLink.
  2. IJ: Ask SW to send a thank-you to the XMLP WG regarding SOAP/GET. See their email to TAG regarding MIME types.
  3. IJ: Ask SW to write thank-you to XMLP WG. See Stuart's email to XMLP WG Chair.

2. Technical

  1. Review SOAP in last call?
  2. xlinkScope-23
  3. Arch Document / httpRange-14

2.1 Review SOAP in last call?

SW: The XMLP WG has asked the TAG whether they will review the SOAP 1.0 specification in last call.

RF, SW: SOAP 1.2 is not sufficiently architectural in scope (in the way the CharMod spec was).

Resolved: The TAG does not intend to commit to a complete review of the last call document. The XMLP WG should indicate whether there is a particular issue having architectural scope they want us to look at.

Action SW: Respond to XMLP WG on behalf of TAG

2.2 xlinkScope-23

Proposal from TBL. In particular, "You should use xlink whenever your application is one of hypertext linking, as xlink functionality such as power to control user interface behavior on link traversal is useful and should be implemented in a standard way to allow interoperability."

I agree with him about that use of hypertext links
RF: I never understood why xlink:href existed. I just use "href" myself.
I am totally amazed at people claiming that human-understandable terms mean something for attribute values
You cannot assume that class, id, href etc happen to mean something!!
href *in the xlink namespace* yes
RF: In the actual documents used by people across a wide range of XML, "href" is consistently used to contain a URI or URI reference." So putting "xlink:href", while it makes us semantically clean in terms of the XML Processing model, the reality is that it doesn't make any difference.
SW: In order for a language to make use of xlink:href, what does it have to import?
CL: You must declare xlink namspace and use namespace mechanism.
<thermostat cref="19C" href="37C"/>
Classic example.
Stupid example. Try using that in practice and users will be upset.
As a concrete example - take an svg file, and declare the "xlink" prefix to a different URL not the XLink one. Your links stops working/
wheras if you declare xmlns:toto fr the XLink namespace URI, it works of course
TBL: If you aren't mixing languages, a namespace defines both syntax and semantics. When you dereference a namespace identifiers, you only get the syntactic bit back.
It's like assuming that an attribute called id is of type ID. It isn't.
TBL: When you mix namespaces, it works with RDF but not XML.
<suit style="double breasted:/>
<social class="proletariat"/>
TBL: With namespaces, if attribute name is not namespace qualified, the attribute is in the element partition on that namespace.
SW: The meaning is scoped by the element that contains it.
CL: And if you find same (unqualified) attrib name on another element, can't say anything about relationship to other (unqualified) attrib name.
RF: I understand the theory, but the practice is that meaning is how it's used.
TBL: Is there software that looks a "href" in generic XML and assumes the value is a URI?
RF: I don't know.
Thats an interesting question, does the element partition have any effect on a namespace qualified name
RF: I think xlink is not limited to hypertext references; it defines relationships in general.
TBL: some questions: (1) Should you bother to use xlink:href, or just invent your own href? (2) If you use a reference but not a hypertext reference, should you use xlink:href?
IJ: Raises the question of the definition of "link" (part of the architecture document). I think link will not be a useful term; context is everything. Too many types of links.
TBL: When we observe that we need the term "link", then we can consider using it.
CL: I think there is a need to identify things that are links: associations that make something part of a whole. There are cases when URIs are just use as disambiguating ids, and these aren't really links. Ultimately the question is whether W3C puts itself behind XLink (for XML grammars).
SW: This is a piece of the mixed language issue.
CL: Yes, you need to do actual web pages, not just backend stuff.
TBL: In the XML Signature world: Take a piece of content, sign it, transmit it, and unsign in another context - this is a function of the target specification. With RDF, you can write down the rules of what happens when you mix statements (meaning of each statement is defined). So within XML, we can't do anything generic. We can talk about special domains (e.g., 2 dimensional graphical rendering and how to mix XHTML, MathML, etc.).
No resolution.

2.3 Architecture document / httpRange-14

Resolved: Publish 1 July draft of arch document.

Action IJ: Publish, asking in particular for input on issue httpRange-14.

  1. ACTION IJ 2002/03/18: Integrate/combine one-page summaries (Revised 1 July)
  2. ACTION TBL 2002/05/05: Negotiate more of IJ time for arch doc
  3. ACTION RF 2002/06/24: Write a paragraph on technical and political aspects of URIs and URI References.


Disagreement over "According to [RFC2396] a resource is "anything that has identity." A resource is part of the Web when there is a URI that identifies it."
TBL: Should be URI Reference. Otherwise you can't refer to some resources.
RF: We are digging a hole for ourselves by saying that abs URis with frag ids define a new space.
RF: URIs point to resources; frag ids are client-side indirect references.
TBL: In RDF, there's another indirection - the frag id is not part of a resource but is about the thing described by the RDF. I have said many times that the phrase "fragment identifier" is a mistake. The meaning is completely language-dependent.
IJ: Are there two self-consistent models here? Which one has advantages?
TBL: Yes, I think you could produce two consistent models.
SW: Seem to be two concepts - identification (speak about it) and dereferencing (get it). Are we conflating those two notions? Would ad-hoc time on the phone help?

3. Postponed

  1. httpRange-14: Need to make progress here to advance in Arch Document.
  2. Internet Media Type registration, consistency of use
    1. ACTION DC: research the bug in the svg diagram. There are two votes to remove the diagram (DC and TB).
    2. ACTION NW 2002/06/24: Produce PNG version of image as well.
  3. uriMediaType-9: Status of negotiation with IETF? See message from DanC.
    TBL: Now there is an RFC for URNs. Sounds like IETF is moving towards URNs and aren't interested in using URIs for media types
  4. Qnames as identifiers
    1. Action NW 2002/06/24: Follow up on Rick Jelliffe comments/proposal.
  5. Consistency of Formatting Property Names, Values, and Semantics
    1. Status of discussion about this finding?
  6. Status of URIEquivalence-15. Relation to Character Model of the Web (chapter 4)? See text from TimBL on URI canonicalization and email from Martin in particular.
  7. RFC3023Charset-21
  8. Status of discussions with WSA WG about SOAP/WSDL/GET/Query strings?
  9. If we get here: namespaceDocument-8
  10. augmentedInfoset-22
    1. ACTION DC 2002/06/17: Talk to XML Schema WG about PSVI. Report to tag, who expects to decide whether to add as an issue next week. Done (email to Schema WG).

3.1 New issues?

  1. Bad practice: Overriding HTTP content-type with a URI reference.See email from TBL.

Ian Jacobs, for TimBL
Last modified: $Date: 2002/07/01 22:46:54 $