SML Teleconference

 

 

Kirk: question about upcoming F2F - what building?

Pratul: will send out information
... start at 1pm on monday, and then 8am tuesday and weds.

<Sandy> Local arrangement page: http://www.w3.org/XML/Group/2007/10/xml-meetings-general.html

Approval of Minutes from Oct 4 2007

resolution: minutes approved

"Schema Binding" Proposal

Kirk: proposal side steps the issue of conflicting schemas. Demonstrates issue with diagram in section 6.

<johnarwe> eg, starting with diagram in 6: remove doc1, posit ref from doc2-v2 to doc2-v1a(or b).

We will hold off the discussion on this until Sandy is present

<johnarwe> kumar: 2 scenarios raised in kumar's email do not appear to be addressed in the newest proposal

<johnarwe> kumar: scenarios are listed in bug 4774

<scribe> ACTION: Kumar will add a simple example to his scenarios [recorded in http://www.w3.org/2007/10/11-sml-minutes.html#action01]

<trackbot-ng> Created ACTION-132 - Will add a simple example to his scenarios [on Kumar Pandit - due 2007-10-18].

<Sandy> section 4.5 is meant to cover the 2 cases Kumar mentioned (and possibly others); 6.4 attemps to answer how it's met.

Kumar: 6.4 doesn't seem to cover the scenarios

Valentina: Would like to have list of all examples that we think are important to be covered by the proposal to ensure that we address them all. Otherwise, examples will continue to come in.
... Everyone please add additional examples now.

Review of Action Items

<johnarwe> http://www.w3.org/2005/06/tracker/sml/users/25462

johnarwe: has action item 119 been completed

<Jordan> http://www.w3.org/2005/06/tracker/sml/actions/overdue

action 119 has been closed

johnarwe: has action 120 been addressed (Refine this proposal to deal with document not reachable due to network issues)

<johnarwe> http://www.w3.org/2005/06/tracker/sml/

johnarwe: we will leave this open

looking at action 97

johnarwe will look at action 97 to determine what to do with it

<scribe> ACTION: 116 to deals with iri's [recorded in http://www.w3.org/2007/10/11-sml-minutes.html#action03]

johnarwe to close 116

kumar: #123 does have a proposal but has not added it to the bug yet

johnarwe: close action item when you add this to the bug

johnarwe: action #5 to be removed

"graph cycles" proposal

Jim: wants to postpone this to another meeting
... recommends that we discuss this at F2F

"keyref" proposal

Kirk: would like to delay this to F2F. Had a conversation with Kumar prior to meeting regarding technical issues with current proposal. Would like to integrate those comments.

<trackbot-ng> Tracking ISSUEs and ACTIONs from http://www.w3.org/2005/06/tracker/sml/

Kirk: appears that either you can allow sml:keyref to refer to xs:key/xs:unique or you can't and have to keep sml symbols space disjoint

johnarwe: any strong need to have this in sml 1.1

Kirk: no

Marv: does anyone else have use cases for this proposal?

no on in group seems to have any

Kirk will rework this, the group will consider whether they have use cases, and the group will discuss this next week at the F2F.

Review and attempt to reach consensus on other non-editorial "hasProposal" bugs, e.g.

<johnarwe> http://www.w3.org/Bugs/Public/show_bug.cgi?id=4978

Valentina: is the proposal to just drop smlerr:output from the SML spec?

Pratul: yes. defining how schematron processors should produce errors belongs to schematron. We don't have structured messages for SML validation itself, so why have this for schematron?
... secondly, it is an optional feature and we have many other higher priority issues to address

Valintina: has microsoft implemented this funtion?

Pratul: no

Valentina: IBM has and she is opposed to remove this. There were others, not only from IBM who wanted this in.

Pratul: Just because something is not in a spec does not mean companies cannot implement. Also, not in charter.
... Appears that this is out of scope given the charter

Valentina: SML error was part of the submission so it is in scope

Pratul: not taking a position just making some points

Kumar: has added issues that he wanted to raise in the bug history

Valentina: suggests the need for a use case and reason to have this in the spec.
... does not believe that the fact that this isn't covered in Schematron doesn't mean that it is not necessary.

Marv: thinks that vijay originally proposed this and would like to hear from him

<scribe> ACTION: Valentina will uncover the use case and reason for having 4978 and will clarify any issues [recorded in http://www.w3.org/2007/10/11-sml-minutes.html#action04]

<trackbot-ng> Created ACTION-133 - Will uncover the use case and reason for having 4978 and will clarify any issues [on Valentina Popescu - due 2007-10-18].

Kumar: is not stating that because something is unclear is should be removed. And so adding clarification is not what he asked for. What he said was that whether or not this gets clarified, it does not belong here.

Valentina: we should try to understand why people thought that it was in scope and she will do the work to clarify the defect

Kumar: if this requires alot of debate it should come after all of the mandatory issues

johnarwe: any issues with moving from "has proposal" to "needs agreement"
... add "needs agreement" and leave "has proposal"

http://www.w3.org/Bugs/Public/show_bug.cgi?id=5095

johnarwe: currently both documents use "interdocument reference" and it means different things

Kumar: is the proposal to use interdocument reference in SML-IF and using cross-document references in SML where it says interdocument
... as long as we define the two terms clearly, he is okay with the proposal. As long as we clarify the terms.

johnarwe: difference between the two terms is as SML uses it, it distinguished between intra and inter document references

Kumar: they are all SML references

johnarwe: SML-IF has a wider definition

Kumar: would prefer not to introduce a new term and instead use sml reference (as opposed to cross document reference)

pratul: would prefer a single clean term. using sml reference in SML would cause less confusion

johnarwe: would not object if we just got rid of inter/intra document reference in SML and just use sml reference
... weed out interdocument references from SML and change to SML reference
... we have consensus

Resolution: consensus on how to address #5095

johnarwe: changing 5095 to editorial

http://www.w3.org/Bugs/Public/show_bug.cgi?id=5098

<johnarwe> 5098

johnarwe: issue is text between heading 4 and 4.1 of the SML spec. Could not see anything in here that makes a normative statement. Proposal is to move that text from where it is to a non-normative section. The end of the introduction is suggested.

Kumar: There is another bug that separates normative and non-normative text. And this could be addressed as part of that larger bug.
... We could leave the text as is and note this issue in bug #5091

johnarwe: will update 5098 to be dependent on #5091, and mark this editorial.

no objection to making 5098 dependent upon 5091

Kirk: moving back to 5095. Points out an issue with interdocument reference use in section 4 of SML spec (bold bullet) called "references" (same text as pasted in to 5095.

johnarwe: points out need for editing

Kumar: clarification on 5098. Address 5091 first, and then address 5098. He believes that as 5091 is addressed then 5098 will be resolved.

Valentina: wants clarification on how the editors should deal with this.

Kumar: if you fix 5098 first, the issue is that none of the sections are marked normative. So this issue should be resolved after the sections are marked.

johnarwe: both 5091 and 5098 are currently "no target"

http://www.w3.org/Bugs/Public/show_bug.cgi?id=5099

<johnarwe> make 5099 dependent upon 5091?

johnarwe: proposal is to make 5099 dependent on 5091 and to mark it editorial

Resolution: agreed

http://www.w3.org/Bugs/Public/show_bug.cgi?id=5104

johnarwe: believes that this is primarily editorial

Kumar: needs clarification on problems

Marv: seems like a best practice, not normative

Kumar: would like this to be normative. Claim is that producer is using URI scheme, but does not use it everywhere. So either you make this claim or you don't.
... more than a guideline because without this there will be inconsistency

johnarwe: you are correct in saying that validation results could be different

Kumar: unless we see harm in keeping it the way it is we should keep it.

johnarwe: the way that is is written is not clear

Kumar: agrees that this is not clear

johnarwe: we did not necessarily agree on what the original intent was.

Kumar: use same bug (#5104) and add clarification

johnarwe: move this to "needs agreement", and will update it to include discussion about two different interpretations (text) of original intent

no additional issues with 5104

http://www.w3.org/Bugs/Public/show_bug.cgi?id=5107

johnarwe: this is redundant text and without the spec one could probably not agree on whether it is or not
... should we leave it to the editors to determine if this is redundant

pratul: believes that this is redundant (reads spec)
... would like the editors to verify this

johnarwe: any objections?

Resolution: move bug to editorial

http://www.w3.org/Bugs/Public/show_bug.cgi?id=5112

Kumar: can we have more time on 5112?

Valentina: seems to be a duplicate of 5169
... SML does not define the rule document. If we fix the defect 5169 then we also solve this one

johnarwe: removing editorial from 5169 and making it dependent on 5112 - any objections?

Pratul: Is the proposal saying that the specific binding mechanism will be defined in SML spec?
... is proposal to define mechanism in SML spec?

johnarwe: concept, not syntax to be defined

<pratul> Section 5 has the following verbiage

<pratul> Model validators that conform to this specification MUST provide a mechanism to support binding of Schematron patterns that are authored in separate documents, i.e., not embedded in schema definition, to a set of documents in a model. The mechanism for binding such Schematron patterns to a set of documents in a model is implementation-dependent and hence outside the scope of this specification.

Valentina: she could not find the definition in the spec

Pratul: believes that the above portion of section 5 does define

<trackbot-ng> Tracking ISSUEs and ACTIONs from http://www.w3.org/2005/06/tracker/sml/

Valentina: it does define but doesn't say that it is the definition

Pratul: the spec defines a rule but not a rule document. May want to deifine this

<johnarwe> A binding is an association of a set of one or more rule documents with a set of zero or more model documents. The documents associated with a given rule document are said to be "bound" to it. For a model to be valid, every document in the model must conform to the constraints defined by every rule document it is bound to. It is permissible for a rule document to have no bindings associated with it, and for a model document to be bound to zero rule documents

johnarwe: first paragraph of SMF-IF 3.5.2 defines a binding.
... further makes and assertion about model validity base on bindings

Pratul: suggests removal of verbage about model validity

johnarwe: thinks this belongs in SML and if we add rule document definition this should cover it

Pratul: doesn't see that binding is covered by a definition of rule document

Valentina: SML spec says that it defines schema documents and rule documents. There is no definition of a rule document in the spec other than this.

Pratul: agrees that we should define rule document in the SML spec
... additional issue is defining bindings
... SML spec has a notion of binding but has no mechanism - this is how bindings must be defined.

johnarwe: syntax stays in SML-IF but that first paragraph (pasted above) should be moved to SML spec
... we have made 5169 dependent on 5112
... the concecptual proposal is to define the concept of binding and rule document in SML and to define the syntax by which rule documents are bound to instances in SML-IF.

no objections to the conceptual proposal

johnarwe: can we move 5112 as presented to "editorial"?

no objections

Resolution: move 5112 to "editorial", make 5169 dependent on 5112, and agreed on conceptual proposal to define concept of binding and rule document in SML spec and the syntax by which rule documents are bound to instances in the SML-IF spec.