W3C

- DRAFT -

SML WG Call December 13, 2007

13 Dec 2007

See also: IRC log

Attendees

Present
Jim, Kirk, Valentina, MSM, johnarwe, MSM.a, +1.408.228.aaaa, ginny, Jordan, pratul, Sandy, [Microsoft], Zulah?
Regrets
MSM, 24th, &, 31st.
Chair
John Arwe & Pratul Dublish
Scribe
Kirk Wilson

Contents


 

agenda at http://lists.w3.org/Archives/Public/public-sml/2007Dec/0107.html

<scribe> scribe: Kirk Wilson

<scribe> scribeNick: Kirk

<MSM> Valentina regrets: 24-31

<ginny> Ginny: regrets 17 and 24-31

John: Raises question of whether we should have calls next week. Pratul to look at minutes

Minutes not yet distributed or people did not have time to review.

No new action ideas or bugs.

4772: Localization

<MSM> http://www.w3.org/Bugs/Public/show_bug.cgi?id=4772

John: Any objection through comment #6?

No objection.

MSM has submitted editorial clarifications, comment #9.

scribe: No objection ot moving these changes to editorial. Editors should look at comment #9.

<MSM> http://www.w3.org/Bugs/Public/show_bug.cgi?id=4992

4992: Object Identity http://www.w3.org/Bugs/Public/show_bug.cgi?id=4992

(MSM: I'll catch up eventually--response time is a little slow for me right now.)

Ginny: Sandy has comments in Comment #20. Mostly editorial changes.

John: Any objection to accepting comment #20 as editorial
... Accept comments in #19 and #20.

Resolution: Move 4992 to Editorial and accept changes in #19 & 20.

5291 http://www.w3.org/Bugs/Public/show_bug.cgi?id=5291

Ginny: Recommends not backing out any of the changes and notes that Kirk recommended a rephrasing of a sentence.

John: Proposal is to accept comment #10 and moved to editorial

<MSM> http://www.w3.org/Bugs/Public/show_bug.cgi?id=5294

Resolution: This is editorial and not needsReview and make change in comment #10.

<johnarwe> resolution above is re: 5291. now beginning to discuss 5294

5294: explanation of smilif:dataType as skip

Resolution: 5294 is Resolved, Closed

<MSM> http://www.w3.org/Bugs/Public/show_bug.cgi?id=5301

5301:

Resolution: Resolved

<MSM> http://www.w3.org/Bugs/Public/show_bug.cgi?id=5318

5318

Resolution: Resolved. Fixed.

4644: Allows assertion on local elements and types

http://www.w3.org/Bugs/Public/show_bug.cgi?id=4644

Sandy's and Kumar's comments. Kumar advises that there are a long number of changes.

Ginny: Did review but did not go into it deeply.

Resolution: Up through comment #5, we will accept. If new issues, there should be new issue opened.

John: Any objection to making comment #6 editorial.

Kumar: is Ok with making it editorial but will discuss with Sandy.

Resolution: 4644 is editorial.

<johnarwe> 5247

<MSM> http://www.w3.org/Bugs/Public/show_bug.cgi?id=5247

5247:

Kumar put in text several days ago.

MSM: Ambiguity in the phrase "not part of interchange set".

Sandy: either it is there or it is not.

MSM: Simpler to require data to have a child and base64 data be non-empty.

John: We need to allow for tombstone documents. If change prevents this, then it will undo other issue resolutions.

MSM: Agrees

Resolution: 5247 is Fixed.

4636 Fragment Idenyifiers.

s/Idenyifers/Identifiers

Kumar: This is just editing some of the words.

Resolution: Fixed.

4774:

http://www.w3.org/Bugs/Public/show_bug.cgi?id=4774

MSM: Will read (starting at comment #14)

Ginny: Suggesting we can step through Sandy's comments in #17.

John: inside 17, subcomment 9: rewording regarding schemaComplete ambiguity.

Ginny: Confusing

Kirk: the name may help to resolve ambiguity.

John: we can resolve after LC.

<johnarwe> ...the part in comment 12 where a rename of the syntax is suggested, not the whole thing

Ginny: name is OK, if we define precisely. We should change definition to reflect Sandy's intention.

<MSM> [I wonder if the description could be "The SML-IF document is not known to be schema-complete".

Resolution: rewording of 9 in 17 is editorial
... point 3 in comment 9 in 17 is editorial

point 12 in comment 9 in 17

Any objecdtion to adding wildcard spec to namespace binding type.

Resolution: No objection.

point 15 in comment 9 in 17

Proposal: have a default="false"

Pratul: Most common case will be schemaComplete="true".

Sandy: explains rationale behind default for schemaComplete="false"

Pratul: Agrees

MSM: Status quo treats it as false; but it doesn't address Pratul's suggestion that this is behavior we want to promote.

Pratul: Withdraws objection.

MSM: Thinks it is better if there were default="true" or make attribute required.

Ginny: Support required an no default attribute

MSM: if default is true, then any incompleteness will be flagged during validation.

<zulah> I prefer default="true"

Kumar; Supports required an no default.

Proposal: adding default and making value 'False"
... Objections from MSM and Zulah
... add default and making value "true"

Ginny and Kirk object

Kumar does not object but prefers third proposal

Proposal: make schemaComplete required and no default value

Resolution: 17.9.15: Make attribute Required with no default value.

<johnarwe> proposed change: steps 6 and 7 are applied in current draft only if schemabindings element is used. should also apply if no such element is present or if the consumer does not support schema bindings

Comment 17.14.1: Schema location--if look at schema binding uses alias. Proposal: same logic is to be used when there is no schema binding or the user chooses not to use it.

<johnarwe> 6 and 7 describe the processing for schemaLocation on include, redefine, xsi:, import

<johnarwe> current draft is somewhat different from proposal. restructuring probably required to be clear when each rule applies.

Kunar: It seems that everyone agrees to the principle; need to be considered by editors.

Resolution: Accept 17.14.1 for section 5.5

Kumar: Processing of redefines/include will be the same.

<johnarwe> so that "no schema bindings" handling of schemalocation agrees with case when schema bindings are there

<johnarwe> ...and the consumer supports them

MSM: Agreement seems to conflict with 3.d and 3.c in current editor's draft.

John: this should be a separate question.
... instruction to editors- Kumar is pointing out that nesting is incorrect is 5.5.

<MSM> [I apologize; I think I led us into a rat-hole. I overlooked the condition in 2.c/d and 3.b/c on whether the IF document is schema-complete or not.

<MSM> ]

17.14.3 for Kumar

John, see comment 16. This says what you want.

scribe: Editorial restructuring is required.

Proposal 17.14.3 go to Editors to resolve. (Editorial)

John: need structural change so as to make sure the case is covered

Zulah: and MSM want review of any proposal

<johnarwe> comments 15 and 16 of 4774 suggest some ways to fix

Resolution: 17.14.3 give to Editorials. Editors will propose solution for review.

4687 Handling DTDs

http://www.w3.org/Bugs/Public/show_bug.cgi?id=4687

Proposal: Allow based64 for all documents, require for those containing DTDs

<johnarwe> current text: There

<johnarwe> can be at most one document embedded

<MSM> s/can be/MUST be/ (twice)

MSM: bottom at point 1:recommends changed to statement: There MUST be at most one.

<johnarwe> msm proposes: There MUST be at most one document embedded

<MSM> also for point 2

Zulah: MSM's comments apply to both points 1 and 2

Resolution: 4687 is Editorial as amended by MSM above.

5091: Distinguish normative vs. non-mornative

http://www.w3.org/Bugs/Public/show_bug.cgi?id=5091

Ginny: These changes assume normative unless explicitly marked non-normative.
... We need to reference the example appendices in text.

Valentina: Original text did not this necessary.
... Original text did not make this necessary

Resolution: We accept proposed 5091 as wholly fixed. Open new bugs for changes.

5303: Recognize SML URI scheme without schema assessment

http://www.w3.org/Bugs/Public/show_bug.cgi?id=5303

Ginny: Question regarding how to process 2 URIs.

Sandy: the SML URI does not flag an error because it might be some other scheme.

MSM: Have we created a situation in which some erroraneous schemes will not be identified.

John: This is a situation because of schemes with overlapping content.
... two editorial comments in comment #1 and comments 4 and 5

Resolution: No objection to making 5303 editorial.

5063: Inheritance of SML Constraints

http://www.w3.org/Bugs/Public/show_bug.cgi?id=5063

John: Discussion on last call.

MSM: Continue to fear complexity to write to enforce this. There is a question of alignment, but it is a question of alignment in a design pattern rather than technical matter.
... Only Type assignment needs to be preserved across restriction. Not a contradiction for us to state that target constraints across restriction, but rules will be hard to write.

<zulah> I can't understand Sandy

<MSM> SG: i feared the complexity, too, but now I support the proposal.

I'm also having trouble following Sandy's discussion.

Sandy: Rules will be hard to read, problem only for implementers.

Kumar: Agrees to with Sandy that implementers could keep alignment with schema.

Ginny: Ok with Kumar's proposal.

<zulah> The issues that SML has with inheritance and assertions are applicable to all Schema users and should be fed into Schema 1.1

Zulah: Issues are general issues with schema inheritance.

<zulah> Much of the design for SML's inheritance is better factored into Schema 1.1

<zulah> This should be sent to the Schema 1.1 group which should address this issue. Not SML

Zulah: This issue should remain open at this point.
... Thinks that it will be difficult for SML to get this correct. Resolution belongs in SML WG.
... Clarifies that this concerns pertains only to this issue.

<pratul> Kirk, Zulah said that "Resolution belongs in XML Schema 1.1 WG"

Zulah: Inheritance issues belongs to Schema WG 1.1
... Issues exist for all schema issues and the work be handed off to Schema WG.

<MSM> [The problem we have labeled 'alignment with XML Schema' is precisely the fact that in XML Schema, constraints on elements contained in complex types are not guaranteed to be preserved across complex type restriction. If XML Schema had a clearer story on that topic, then the task of SML in describing this constraint on the target* constraints would be far simpler.]

Zulah: Objects to closing this particular bug and would like to have more discussion.

Pratul: We don't have a dependence on Schema 1.1.

Kumar: This will keep our group hanging for a long time.

John: Strawpoll. Who would object closing this bug as Kumar proposed per sandy's comment in #7?

MSM: there is an architectural issue here. It is premature to close this bug at this time.

Jim: No problem with statement in comment #6. We need to come up with a schedule for closing this and plan.

<Jim> I suggest we not close it without entertaining the possibility that another option be taken.

Ginny: We need to support the request not to resolve this issue due to objection from Zulah and MSM.

Jim: HP does not have issues as laid out by Kumar.

Kumar: It doesn't make sense to have issues addressed by Schema. 1.1 WG
... What is technical issue?

Zulah: Technical issue is that we need to consider architectural issue. These are larger than SML WG.

Kumar: What IS the architectural issue here?

<zulah> We believe that going to LC with will slow the progress of SML

<MSM> Zulah: the technical issue is the one identified in 5063: inheritance of constraints on element declarations across complex type restriction

Kumar: We are not changing XML Schema constraints, this is an issue of SML constraints that are not handled by Schema 1.1 WG.

Zulah: BEA's contention, if Schema 1.1 addressed these issues, then SML wouldn't have to.
... BEA does not yet have a proposal at this point. It needs more discussion.

Kumar: This is not an Schema 1.1 architectural issue.

MSM: Complex type restriction propagate across restriction, but contraints on child element declarations do not propagate.
... In 5063, we want our constraints to behave different.
... Therefore, Schema has gotten it wrong.

I have lost telephone connection. Can someone continue recording this the discussion?

<johnarwe> will do

<johnarwe> time to end anyway

I can't get back on because the "conference is restricted"

<johnarwe> MSM: note that when doing derivation by restriction in schema, constraints on types are automatically inherited and constraints on elements are not automatically inherited.

Let me know when I can generate the minutes.

Is the discussion still going on?

<johnarwe> next meeting Jan 3

<johnarwe> request to BEA to email some details on their issue, including a proposal, while people are out next 3 weeks.

<johnarwe> Zulah notes BEA has forced closure next several weeks, so may need to wait several weeks into Jan for a thorough proposal.

rssagent, generate minutes

rssagent, make log public

see my email that I just sent. The issue ended immediate after you hung up.

I meant "session" rather than "issue". Have "issue on my mind.

Ok Now it's working.

And I was able to access the minutes--WHEW!! That really scared me.

I don't think the rrsagent likes me :-)

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.128 (CVS log)
$Date: 2007/12/13 21:37:35 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.128  of Date: 2007/02/23 21:38:13  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/Any/No/
FAILED: s/Idenyifers/Identifiers/
Succeeded: s/required/required an no default/
Succeeded: s/false/true/
WARNING: Bad s/// command: s/can be/MUST be/ (twice)
Succeeded: s/and have/and would like to have/
Found Scribe: Kirk Wilson
Found ScribeNick: Kirk
Default Present: Jim, Kirk, Valentina, MSM, johnarwe, MSM.a, +1.408.228.aaaa, ginny, Jordan, pratul, Sandy, [Microsoft], Zulah?
Present: Jim Kirk Valentina MSM johnarwe MSM.a +1.408.228.aaaa ginny Jordan pratul Sandy [Microsoft] Zulah?
Regrets: MSM 24th & 31st.
Got date from IRC log name: 13 Dec 2007
Guessing minutes URL: http://www.w3.org/2007/12/13-sml-minutes.html
People with action items: 

[End of scribe.perl diagnostic output]