SML WG f2f in Redmond

28 Oct 2008

See also: IRC log


John, Kumar, Ginny, Sandy, Len, MSM, Kirk, Pratul
Jim, Julia
Kumar, Ginny



Test samples below left over from previous day's IRC log, due to time zone differential between UTC and PT.

<MSM> XSLT test summary: http://www.w3.org/XML/Group/xslt20-test/Documentation/reportSummary.html

<MSM> XSLT test cases detail: http://www.w3.org/XML/Group/xslt20-test/Documentation/reportTestCases.html

<MSM> XQuery implementation report: http://www.w3.org/XML/Query/test-suite/XQTSReport.html

EPR note

kirk: I will walk you through the doc.

<johnarwe_> http://lists.w3.org/Archives/Public/public-sml/2008Oct/0016.html

kirk: ...explains the need to have EPR based SML ref scheme.
... ... section 2 defines a framework for defining EPR based SML ref schemes.
... ... discusses how section 2 provides a way to convert the non-normative note into a normative one if desired.
... ... describes the sample EPR based scheme defined in section 3.1.
... ... describes how SML refs that use EPR based ref schemes can be used in SML-IF for interchange.

msm: Publishing first public draft of a note is relatively easy. The general understanding is that the WG is done with the content of the note and is not expected to work on it more.
... publishing as a working draft ==> the WG will work on it further.
... publishing as a note ==> WG does not expect further work on it.

kirk: I propose that we publish the EPR note as a W3C note.

RESOLUTION: The WG agrees that the EPR note will be published as a note.

Discussion on how we treat model processors.

john: we define model processors in the spec and there are a few normative statements about model processors in the sml spec.
... The definition of model processors is so broad that we cannot have any meaningful set of tests to test interoperability of model processors.

msm: we have some statements in SML spec that use MAY or MUST regarding model processors therefore we should add an entry in the "Conformance" section for model processors.

<johnarwe_> A conforming [model processor] MUST process a conforming SML model using, in whole or part, semantics defined by this specification. It is OPTIONAL that a conforming model processor process all elements defined in this specification, but any element that is processed MUST be processed according to the requirements stated in the normative sections of this specification. In particular, if a conforming model processor performs model validation, then that proc

<johnarwe_> In particular, if a conforming model processor performs model validation, then that process MUST be performed as described in this specification.

<johnarwe_> further, ginny suggests omitting that last sentence

<johnarwe_> MSM: we have established that both specs have must/may/should statements which target model processors in general

<MSM> A conforming SML [or: SML-IF] processor is one which satisfies all the

<MSM> constraints imposed on processors elsewhere in this specification.

<MSM> Optionally, add: All SML[-IF] validators are SML[-IF] processors, but

<MSM> not all processors are validators.

RESOLUTION: Add the first two lines proposed by MSM to the SML Conformance Section. [That is, "A conforming SML [or: SML-IF] processor is one which satisfies all the constraints imposed on processors elsewhere in this specification."]

<john> (post-meeting) now reflected in bug 6205

Continue discussion of SML transition request document

<johnarwe_> after we finish with the EPR Note review, and pick up the transition request again, the drafts reflecting yesterday's wg mtg updates are at http://lists.w3.org/Archives/Public/public-sml/2008Oct/0020.html

ginny: Isn't multiple schemes in an SML reference more of an extension point rather than optional feature?

MSM: propose that we label this feature as EXT not SML

Kumar: we should add additional info about our categories of optional features, e.g. testability of extension points.
[test case] need test case for non-Schema determined IDs
... we can't test whether an implementation does consistency checking on multiple base uri methods

MSM: if inconsistency is known, the model is invalid

Is there a requirement to report this?

MSM: 'xml:base wins' means we prescribe what interpretation is to be placed on the model
... you recover from this error (inconsistency) this is how you recover.

[test case] a test case for consistency checking assuming that we have an implementation that does consistency checking.

No candidates for "at risk" identified in SML

No candidates for "at risk" identified in SML-IF

the working group will define a list of features and verify implementation of these features in the 2 implementations we have

RESOLUTION: The WG agrees that we are ready to request transitioning the specifications to the CR stage.

Creating list of features

Discussion of granularity of features

spec table of contents is the starting point

List is being captured in the SML transition request doc.

Initial feature list identified; now working on matching Cosmos tests with features.

Summary of Action Items

[End of minutes]

Last Scribe Date        Member Name         Regrets pending 
2008-05-22              Lynn, James         Until further notice
2008-07-10              McCarthy, Julia     Until further notice
2008-xx-xx              Charest, Len	
2008-10-16              Wilson, Kirk
2008-10-27              Gao, Sandy
2008-10-28              Pandit, Kumar
2008-10-28              Smith, Virginia 
Exempt                  Arwe, John 
Exempt                  Dublish, Pratul
Exempt                  MSM