See also: IRC log
RESOLUTION: approve 2008-11-13 telecon meeting minutes at http://lists.w3.org/Archives/Public/public-sml/2008Nov/att-0029/20081113-sml-minutes.html
Action 203 is done.
RESOLUTION: these calls are cancelled.
Kumar: expected publication date Nov. 25.
Kumar: namespace policy was not included in the publication request.
Kumar: should we have one for each namespace uri, or one that covers all the namespaces?
John: seems we need one for each namespace URI, which is the same as Len's reading.
Ginny: how about one document but linked from multiple places?
Kumar: will confirm with W3C whether it's OK to have single doc with multiple links.
All: OK with Kumar's draft.
RESOLUTION: Kumar to consult with W3C with his draft.
Len: now has CVS setup OK. wondering where the XLink note should go.
John: likely to put the notes in the same directory as the specs.
<ginny> [Public] / 2007 / xml / sml
Kumar: source/build files need to go to "sml" directory.
<ginny> http://dev.w3.org/cvsweb/2007/xml/sml/
John: seems 4633 is done in the CR candidate draft. should be closed.
Kumar: yes. I will close 4633.
John: 6207 is for CR changes. should also be closed.
<johnarwe_> http://www.w3.org/Bugs/Public/show_bug.cgi?id=6237 is new, found looking at cosmos code
John: can documentURI be
relative? e.g. to refer to documents in the same directory on
the file system.
... spec doesn't answer one way or the other
Ginny: sounds like relative should be allowed.
John: suggest to use "should". That is, contents of documentURI elements *should* be a URI reference.
<johnarwe_> http://www.w3.org/TR/2004/REC-xmlschema-2-20041028/#anyURI
Ginny: the schema type is anyURI, so relative references are already allowed by the normative schema.
Sandy: then what base URI should be used to resolve such relative references?
<johnarwe_> http://www.w3.org/TR/sml-if/#Base_URI
John: 5.3.4 is the relevant section about where the [base URI] is applied.
<johnarwe_> http://www.w3.org/Bugs/Public/show_bug.cgi?id=6191
<johnarwe_> 11/6 telecon agreed on the following revision.
<johnarwe_> smlif 5.4.2, betw existing parag 2 and 3, add:
<johnarwe_> Note: Since the URI prefixes specified as ruleAlias and documentAlias elements
<johnarwe_> are aliases, they are subject to all of the processing for aliases as described
<johnarwe_> in [5.3.3]. For example, if they are relative references then they would be
<johnarwe_> transformed to absolute URIs before comparison.
John: at 2008-11-06 telecon, we decided that 5.3.4 does cover aliases. A non-normative note was added. See above under bug 6191.
<johnarwe_> It sounds like the proposal emerging is, in section 5.2.2 of SMLIF where it currently says: SML-IF specifies one way that MAY be used to provide the location of the referenced document, the documentURI element.
<johnarwe_> Have it say: SML-IF specifies one way that MAY be used to provide the location of the referenced document, the documentURI element. (same so far)
<johnarwe_> ... (new) It is a consequence of documentURI's schema definition that it contains a URI reference, i.e that it may be an absolute URI or relative reference. When it is a relative reference, the [base URI] property SHOULD be used to transform it to an absolute URI, as stated in [5.3.2 Base URIs].
John: we can add a "should" to say: if documentURI value is relative, then [base URI] *should* be used to resolve it. See above for details.
Ginny: this matches my expectation.
RESOLUTION: adopt the above wording change to resolve bug 6237. No further review needed.
<scribe> ACTION: Michael to add "PR" as a valid "target milestone" value in SML bugzilla system. [recorded in http://www.w3.org/2008/11/20-sml-minutes.html#action01]
<trackbot> Created ACTION-206 - Add \"PR\" as a valid \"target milestone\" value in SML bugzilla system. [on Michael Sperberg-McQueen - due 2008-11-27].
John: anyone thinks the impl reports contain member-confidential material?
All: no.
RESOLUTION: move impl report discussion to the public mailing list.
Kumar: about the impl report format, I think it's easier if the source is in something like Excel, then transform to xml.
John: think we are OK with using Excel for now, however we need to make it accessible to those without that app (text, html, xml, something like that) now as we go, and at the end one of these forms (presumably xml) becomes the authoritative copy so it is easy for others to pick up later if need be. Wondering how to merge changes made by different people.
Kumar: we can try to merge the xml file and import back into Excel.
John: Pratul raised concerns several weeks ago about using the generic link (latest from HEAD) to the COSMOS CVS repository. We may need to point to specific versions of the test cases.
Kumar: better to use the generic
pointer (to the latest version) now. after things stabilize, we
can go through them and add version numbers.
... COSMOS impl report looks OK.
Ginny: also looked at the diff for the COSMOS report. what's the next step?
John: need to look at the
"potential problem" ones and decide what to do.
... first such instance is "SchemaBindingNS". If schema binding
is supported, then "SchemaBindingNS" can't be implemented, yet the latter is listed as Required.
Kumar: sounds like we need a new "conditionally required" category, in addition to "required" and "optional". The condition can be included in the "description" field.
John: Same problem with "smlifBaseUri" and "locatorNS"
Kumar: I will add the new "CR = conditionally required" category. Will import changes made to description and implementedByCosmos elements.
John: noSchemaNV. no implementation for this. should reconsider where this feature belongs. suggest to move from "interop" list to "extensibility" list.
RESOLUTION: move "noSchemaNV" from "interop" list to "extensibility" list.
John: Similarly, should "baseUriCheck" be an interop feature?
Kumar: doesn't look interop.
Ginny: agreed.
RESOLUTION: move "baseUriCheck" from "interop" list to "extensibility" list.
John: "baseUriReq" seems to be a duplicate of "baseUri" (renamed to "baseUriIfRelativeURIExists" in COSMOS report)
<johnarwe_> old/new line numbers: 105/109 smlif:baseuri 139/152 xml:base 163/176 at least one of them required 219/237 appears to dup the prev one(s)
Ginny: agree that "baseUriReq" is a duplicate of "baseUri"
RESOLUTION: remove "baseUriReq" from the feature list.
Kumar: I will make changes agreed upon today and send out the master copy of the source.
John: will go through Ginny's
list and see whether COSMOS already has tests to cover them.
e.g. non-schema-determined ID.
... should we put the impl report, feature list etc. at a fixed
location, so that we don't have to fish through the archive all
the time.
Ginny: sounds like a good idea.
Kumar: I can add Ginny's list to my Excel. That'll make it easier to manage the lists.
<scribe> ACTION: John to put content of Kumar's zip file at a location, preferrably publicly accessible. [recorded in http://www.w3.org/2008/11/20-sml-minutes.html#action02]
<trackbot> Created ACTION-207 - Put content of Kumar's zip file at a location, preferrably publicly accessible. [on John Arwe - due 2008-11-27].
Last Scribe Date Member Name Regrets pending 2008-05-22 Lynn, James Until further notice 2008-10-29 Charest, Len 2008-10-29 Pandit, Kumar 12/4 - 1/8 inclusive 2008-11-06 Wilson, Kirk 12/4 2008-11-13 Smith, Virginia 12/18 2008-11-20 Gao, Sandy Exempt Arwe, John Exempt MSM