SML Conference Call, March 20, 2008

 

 

<scribe> scribe: Kirk Wilson

<scribe> scribeNick: Kirk

<johnarwe> regrets from zulah

<johnarwe> regrets for next week from Jordan

Approval of minutes

<johnarwe> http://lists.w3.org/Archives/Public/public-sml/2008Mar/att-0075/2008-03-13-sml-minutes.html

No changes; No objections

Minutes are approved

Message from the CG: F2F meetings after June

John: Think about this, perhaps talk about it next week. CG will schedule another meeting in October around TPAC.
... We will be better able to see what we need to do after we get a public comments. We'll see if we need another Last Call.

Jim: What is schedule if we don't get substantial comments?

John: It's been a while since I read this part of the Process, my memory says minimum is 3 to 6 months. CR in June/July, Proposed by October. We also need to do interop testing and reporting around that. PR is Sept/Oct.
... We are expecting 2 implementations: Microsoft and COSMOS (IBM).

MSM: How long will these implementations take?

Kumar: Implementation exists; need to update for changes in LC. Not a long CR period.

John: COSMOS just needs to do upgrade from last call. Expected end of May

John and MSM discuss typical needs of the group in moving forward for CR.

John: Think about what we might need. Valid to say no more F2Fs.

Action Items

Pratul: Has not looked at 137.

Kumar: 171 working on, no proposal yet

Kumar: 175: email thread on 5283 ("interchange set" definition)

MSM: add users to Bugzilla. Ability to cc users to enter bugs. Believes he did this; but can't tell for sure. MSM will monitor.

Pratul: only one other person for XLink.

MSM: Will close action item.

No key word or Target

5543: http://www.w3.org/Bugs/Public/show_bug.cgi?id=5543 Comment from Henry

Kumar has comment

Kumar: Constraining media type is permitted in an XPointer scheme.
... If you use xpath1 scheme id="abstract" would get you to the same element.

MSM: Not allowing bare-names lacks a good motivation.
... MSM is not certain that he is happy with the response.

Sandy: Is what we are doing allowed? / Rather it will help people to process bare-name schemes, (3) Are we restricting our scheme to allow only the this scheme.

Pratul: Henry wants discussion in public. It is not clear what he wants.

MSM: Post response in Bugzilla and ask for clarification of what the issue is.

Pratul: We should add note to bug.

Kumar: We need to word it politely to ask for response.

MSM: We need to understand the hypothesis involved in the issue and respond those. Not allowing bare names will cause confusion. Specifying only a subset of URIs to be used is confusing. We need to show that our restrictions do not violate architecture of the Web.
... We need to understand that many Web people do not understand the rationale of SML. We need to get our story across better. Saying that "you're not clear enough" is not good enough. It indicates that our spec does not explain how it fits into larger picture.

Kumar: Agrees with this approach. Main question: Should we allow any URI. Kumar wrote response around this question: answer relates to interoperability.
... Difference between dereference SML Ref in SML-IF vs. derefencing locator element of a non-embedded document (on which you do alias matching).

Sandy: two senses of "dereference": talking to the protocols vs. redirection to a cache, etc.

<johnarwe> At the risk of fulfilling my role as co-chair, it's probably time to say 5543 is needsagreement and continue triage

Kumar: In context SML-IF, we just do prefix matching. Not caching, etc. We don't actually execute the protocol.

John: Mark this bug as needsAgreement. No objection.
... We will continue triage and later draft someone to draft response.

5544: http://www.w3.org/Bugs/Public/show_bug.cgi?id=5544

MSM: Kumar's response just seems to restate the question. Constraints of SML arose from a specific domain. But we have split those constraints off for a general approach. Maybe we now want to point to something other types of things--e.g., images. We wanted a more general purpose validation mechanisms not tied to data center ops.

Pratul: Our considerations are relevant only when target is documents.
... We would have to revise definition of a model !! Whole document would need to change.

<pratul> Kirk: Kumar made the comment about "whole document would need to change"

MSM: Agrees, targetType would be more elaborate and complicated. See W3C link checker.

Discussion of what targetType means in this case.

scribe: Impact on notion of model validation is an issue.

Kumar: It is not a good tradeoff. We could add a note saying that targetType/targetRequired could be used in other applications, but "not a good thing" in this spec.

MSM: Questions: what is "not good"?

Ginny: even if apply targetType to XML documents, definition of model would have to change.

John: Mark this bug as needsAgreement. No objection.

5545: http://www.w3.org/Bugs/Public/show_bug.cgi?id=5545

Kumar: 5545 depends on 5508.

5546: http://www.w3.org/Bugs/Public/show_bug.cgi?id=5546 (Why need SML-IF?)

John: Another case where our design rationale is not clear.
... Mark bug as needsAgreement. No objection.

5558: http://www.w3.org/Bugs/Public/show_bug.cgi?id=5558

Discussion Kumar/Ginny regarding how to resolve.

Ginny: Definition of document needs to remove "well-formed XML document", because that's the conformance criterion.

MSM: What is this rule trying to catch. Error is that a document is not a well formed XM document.

<MSM> <sml:document href="http://www.w3.org/Icons/w3chome.jpg"/>

MSM: Things pointed to are well-formed documents. The JPG is not a document in the model.

John: ADMONITION TO GROUP: We need to get a couple of these issues fleshed out between calls, so they can be resolved next week. Let's not fall into the trap of ignoring them between calls, instead proceed on a couple via email.
... Mark issue as needsAgreement. No objection.

5561: http://www.w3.org/Bugs/Public/show_bug.cgi?id=5561 (Define other schemes)

Ginny: We should define a new reference scheme; it would be a good "exercise", but should not be included in the text.

Kumar: How would XHTML be different from what we are doing?

Pratul: Henry wants to use existing documents and treat them as SML references.
... We don't recognize these as SML references.

Ginny: We should prove it is simple, but should not present it.

<pratul> The following text is from a comment made by Henry Thompson

<pratul> In my opinion, not good enough. The SML spec. itself should define both an

<pratul> XHTML href Reference Scheme and a Simple XLink Reference Scheme. Either it's

<pratul> easy to do this, so you definitely should, or it's hard, in which case that

<pratul> uncovers a weakness in your spec.

<MSM> http://www.w3.org/Bugs/Public/show_bug.cgi?id=5513

John: It sounds like Henry thinks it should be a normative part of the spec. (Ginny agrees.)

MSM: We could say this is a discussion for a separate Note (like for EPR).
... Vision for SML: Have application beyond original domain, and we should "seize" the new opportunity. E.g., Checking links, using other vocabularies, such as XLink.
... Need to recognize reference schemes in HTML documents, etc. We have potential to define a technology can that do something interesting with respect to existing document.

Pratul: We don't have expertise, not in our scope. Other groups should do this given that we have defined what it is to define a ref. scheme.

MSM: Less involved understanding hypertext links than involved in understanding SML.
... We are in better position to demonstrate usefullness.

Pratul: Disagrees with MSM's position.

MSM: How is XHTML domain specific???

Pratul: Used in rendering info in browser.

Ginny: Go throught the exercise and reserve judgment on actually publishing it.

John: Ginny's Proposal: Take at least one of schemes in 5561 & 5562: we should go through a "paper" exercise and reserve judgment to see if it should be published.

Kumar Objects.

John: Mark bugs (5561, 5562) as needsAgreement
... 5564 (we have already discussed it.)

5580: http://www.w3.org/Bugs/Public/show_bug.cgi?id=5580 (change in definition of URI scale)

John: Does proposal reflect will of group?

Kumar: 4632 has resolved the issue. Explains history. But is about restoring original definition.
... See MSM's comment #3 in 4632.

Sandy: Asks for more time to review the issue.

John: if we go with MSM's comment, is there a consistency problem with ref. to RFC 3986?

MSM: 3986 is a tighter statement of the other RFCs. Checking URI seldom is affected by the subtleties of these RFCs.

Pratul: This is editorial error. (Kumar agrees.) Why are we debating this?

John: Question is whether this changes any implementations. This has process implications.

Kumar: To his knowledge it does not make any implementations invalid.

Sandy: Do we care about other references to RFCs?
... Wants more time. We need to be consistent.

Meeting adjourned at 4:03

Last Scribe Date  Member Name               Regrets pending 
2008-01-22        Eckert, Zulah 
2008-01-23        Smith, Virginia 
2008-02-07        Lynn, James 
2008-02-14        McCarthy, Julia 
2008-02-21        Kumar, Pandit 
2008-03-06        Boucher, Jordan              03/27/08 
2008-03-13        Gao, Sandy 
2008-03-20        Wilson, Kirk 
Exempt            Arwe, John 
Exempt            Dublish, Pratul 
Exempt            MSM 
Exempt            PH