SML Face-to-Face -- 12 Jun 2007

<Kirk> scribe: Kirk

who is present?

Discussion of section 2.2

who is present?

Philippe: explained use of "@"s in XML Namespace table in 2.3

<MSM> In the room (let's see, it's a test of whether we learned people's names), I see: Pratul Dublish, Vijay Tewari, Paul Lipton, Kirk Wilson, Marv Waschke, John Arwe, Sandy Gao, Ginny Smith, Jim Lynn, Philippe Le Hegaret, Michael Sperberg-McQueen, Valentina Popescu, Bassam Tabbara

<plh> Philippe: you might want to consider using http://www.w3.org/ns/sml when you get to CR

<Paul_L> present

<scribe> ACTION: WG consider switching to short namespace [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action01]

<trackbot-ng> Sorry, couldn't find user - WG

<Paul_L> who is here

<plh> ACTION: John to consider switching to short namespace at CR [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action02]

<trackbot-ng> Created ACTION-6 - Consider switching to short namespace at CR [on John Arwe - due 2007-06-19].

<plh> URIs for W3C Namespaces

Michael: Extension of Identity Constraints is different than referencing

Sandy: Clarify final sentence in section 3, use of "extension"

Vijay: suggests saying "SML uses to points of extensibility of XML Schema"

Sandy: recommends reference to specifc suggestion

<MSM> Proposal for final para of section:

<MSM> XML Schema's validation semantics can be extended by schema

<MSM> annotations in the form either of attributes in other namespaces or of

<MSM> xsd:annotation elements (sec. n.n of [XML Schema: Structures]; both

<MSM> forms are used by SML.

<Sandy> Sandy suggests to explicitly mention "annotation" with a reference to the schema spec.

<Sandy> http://www.w3.org/TR/2004/REC-xmlschema-1-20041028/#cAnnotations

Vijay: are we really extending schema validation semantics?

Michael: Yes--What SML is doing does classify as extending validation semantics

<MSM> Alternative 3:

<MSM> SML extends XML Schema 1.0 by schema annotations both in the form of

<MSM> attributes in other namespaces and of xsd:annotation elements

<MSM> (sec. 3.13 of [XML Schema: Structures]).

johnawre: Exact words should be editor's job

Ginny: would like to everyone to speak up on this sentence

Marv: To resolve issues quickly, editors should go to the people who feel most deeply.

Michael: If editors change the text to improve, they should "due diligence" and consult with those who feel deeply.

Section 3.1 (XML Schema Profile)

Sandy: Questions phrase "SML supports a strict subset of XML Schema" but "extends" it.

<scribe> ACTION: Bassam to remove first sentence of the section. [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action03]

<trackbot-ng> Created ACTION-7 - Remove first sentence of the section 3.1. [on Bassam Tabbara - due 2007-06-19].

rrasgent, generate minutes

<Paul> rssagent, here

<johnarwe> A good place to bookmark as an anchor to find all the bot commands etc is http://www.w3.org/Guide/

Michael: Subsetting of 1.0 seems orthaganol to rest of spec. Seems like a "style guide".
... Propose that we eliminate the idea of subset.

Bassam: In CML, there is a place for guidelines. This sort of thing is appropriate there rather than here.

Pratul: Nothing in spec depends on this.

Michael: SGML experience raises concern about subsetting spec; raises interoperability issues.
... If the section is a problem it should be sent to XML Schema WG.

Bassam: Proposes that we treat specification as a superset of XML Schema. Don't talk about "profile" of XML Schema. Guideless are treated in such areas as CML.

Johnawre: Avoid xs:redefine in order to preserve backward compatility.

<ginny> testing

Discussion between Paul and Bassam on value of preserving ideas of section in an appendix. Bassam: this depends on context of use of the feature.

<scribe> ACTION: Bassam to remove section 3.1. (This supersedes ACTION-7.) [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action04]

<trackbot-ng> Created ACTION-8 - Remove section 3.1. (This supersedes ACTION-7.) [on Bassam Tabbara - due 2007-06-19].

<Paul> Paul: Regarding discussion of 3.1, I believe it important that SML have no references to any particular group working on the instantiation of a particular model.

<MSM> trackbot, op me

<MSM> no.

Section 3.2

Michael: Proposition 1 Cross Validation of documents = putting documents in one document.
... Proposition 2, there are things you can do in SML that cannot do under Proposition 1

Ginny: No experience contrary to Proposition 1.

Bassam: richer set of constraints around identity in SML references.

Michael: The richer constraints might be taken up by XML Schema group, but right now the answer to his concern is in favor of Proposition 2.

Jim: raises question on whether model needs to be a tree. Bassam: Nothin requires this.

johnaware: Is anyone to be liaison with XML Schema WG.

Discussion of this group's relationship to the XML Schema WG, esp. RE the point that xs:redefine should be dropped from XML Schema.

<scribe> ACTION: Johnawre to present best practices as defined by SML domain groups to XML SChema WG for their consideration. [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action05]

<trackbot-ng> Sorry, couldn't find user - Johnawre

<scribe> ACTION: Johnarwe to present best practices as defined by SML domain groups to XML SChema WG for their consideration. [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action06]

<trackbot-ng> Sorry, couldn't find user - Johnarwe

<scribe> ACTION: john to present best practices as defined by SML domain groups to XML SChema WG for their consideration. [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action07]

<trackbot-ng> Created ACTION-9 - Present best practices as defined by SML domain groups to XML SChema WG for their consideration. [on John Arwe - due 2007-06-19].

<plh> trackbot-ng, status

Discussion of section 3.2

Philippe: Consider using IRIs rather than URIs.

Pratul: Should we remove discussion of URIs in favor of IRI or add IR.

<scribe> ACTION: Philippe to put issue on this in Bugzilla. [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action08]

<trackbot-ng> Created ACTION-10 - Put issue on this in Bugzilla. [on Philippe Le HÃ©garet - due 2007-06-19].

Philippe: Raises issue of use of ReferenceParameters in EPR to reference a resource.

Vijay: The issue is how to construct a EPR to a subpart of a document.

<plh> Using abstract properties of an EPR other than [destination] to identify resources is contrary to this recommendation.

Paul: Suggests using XPOINTER expression in the wsa:Address to identify the subpart of the document.

Vijay: nothing architectural to guarantee that the URI will identify that part of the document.
... We need a mechanism to define how to architecturally define document parts by EPR as URI.

<Vijay> Reference to Section 2.6 of WSA http://www.w3.org/TR/ws-addr-core/#resourceidentification

<scribe> ACTION: Bassam to remove ReferenceParameters and change wsa:Address to match that sml:uri in the element above in the example. [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action09]

<trackbot-ng> Created ACTION-11 - Remove ReferenceParameters and change wsa:Address to match that sml:uri in the element above in the example. [on Bassam Tabbara - due 2007-06-19].

ACTION- 11

Discussion of 3.2.1

<Paul> As per request of scribe, modified action 11. Now says:"Bassam to make sml:uri tag have plain URI and do the same in the EPR example in the address tag. removing EPR reference parameters."

John: We need to cover in 3.2.1.1 situation in which one element with sml:ref="true" with multiple occurrences of the same scheme.
... all references must point to the same thing.

Zulah existed meeting (before discussion of 3.2).

<Paul> s /existed/exited/

Discussion of this issue.
...Bassam: Spirit is that there is one reference. Resolution of multiple references is implementation dependent; modify second sentence. of 3.2.1.1
...Pratul: a single reference may have multiple representations.
... they must target the same reference. Checking this optional.

<scribe> ACTION: Bassam to change second paragraph of 3.2.1.1 to "If a single reference has multiple representations, every representation must target the same element. Validators may check this condition." [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action10]

<trackbot-ng> Created ACTION-12 - Change second paragraph of 3.2.1.1 to \"If a single reference has multiple representations, every representation must target the same element. Validators may check this condition.\" [on Bassam Tabbara - due 2007-06-19].

TOPIC . Discussion of 3.2.1.2

<scribe> ACTION: Bassam to remove the word "different" from the first sentence of 3.2.1.2. [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action11]

<trackbot-ng> Created ACTION-13 - Remove the word \"different\" from the first sentence of 3.2.1.2. [on Bassam Tabbara - due 2007-06-19].

Discussion3.2.1.3

Sandy: There may schemes for representing references that use ony attributes. Therefore, xsi.nil="true" would be ineffectual in this case.

Pratul: The spec specifies (section 3.2), that the reference must be a child element.

Bassam: agrees with Sandy's point that spec leaves this issue open for new schemes that defines the address in the attribute.

<scribe> ACTION: Sandy to open an issue to address whether we should require reference pointers be children. [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action12]

<trackbot-ng> Created ACTION-14 - Open an issue to address whether we should require reference pointers be children. [on Sandy Gao - due 2007-06-19].

<MSM> so we need to decide: do we want to require that pointers be given in child elements? If so, we need to say so more explicitly; if not, we need to revisit the story about nil in 3.2.1.3

Valentina: Question on John's issue and use of IRI-schema--should be change sml:uri to something more general?

<plh> Use of IRIs

Sandy: Difference xsi.nil="true" between a "dangling" reference?

Bassam: difference is that nil means there is a no address. Dangling reference has not syntactic representation.

<BassamT> ACTION: Bassam to correct section 3.4 targetRequired such that targetRequired=true is not violated for empty or null references. [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action13]

<trackbot-ng> Created ACTION-15 - Correct section 3.4 targetRequired such that targetRequired=true is not violated for empty or null references. [on Bassam Tabbara - due 2007-06-19].

Discussion of difference between NULL (nil="true") and Empty.

Michael: Nil="true" means there is no children. Therefore, restrict not checking reference to just nil=true. Empty elements with nil=false gets treated as a reference.

Pratul: We have proposal from Michael to consider to his point in the text.
... Ginny points out that the element may have content but not a reference. This supports Michae's approach.

Jim: question: What is ignored if nil="true"?
...John: xsi.nil=true is equivalent to sml:ref="false".

<scribe> ACTION: Bassam to chancge second sentence of 3.2.13. "A model validator MUST NOTdo the SML validation associated with references for that element." [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action14]

<trackbot-ng> Created ACTION-16 - Chancge second sentence of 3.2.13. \"A model validator MUST NOTdo the SML validation associated with references for that element.\" [on Bassam Tabbara - due 2007-06-19].

John: No where in the spec do we have a definition of a dangling reference.

<Paul> s /Chancge/Change/

<scribe> ACTION: John to open Bugzilla issue to address defifnition of a dangling reference. [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action15]

<trackbot-ng> Created ACTION-17 - Open Bugzilla issue to address defifnition of a dangling reference. [on John Arwe - due 2007-06-19].

Paul: raises question of how to use tools--when do we make issue an Action vs. goes into Bugzilla.
...John: Public use may involve detailed "bugs" on highly specific issue.

<Paul> Paul: suggests that use of bugzilla vs. making something an action item is based on the attention of this WG. If the WG in real time is resolving something now, it is an AI. If the WG defers something or something is brought to the attention of the WG from a public source, for example, it goes into bugzilla. Bugzilla, in short, is a mechanism to place things in a "queue" for later consideration by WG.

Vijay leaves meeting at lunch.

Following an excellent lunch provided by Microsoft: Paul presents his suggestion, above, for group to agree up.
...Sandy: XML Schema group tracks all issues in Bugzilla for tracking.
...Kirk: at point of Last call, we need to track every change to the spec so that the Director can review changes.
...Michael: this can

Paul: Issue is were the source of truth? In which system.
... is the truth?

Zulah returns to meeting.

<Paul> I said: The basic real time meeting activity of this group, in a sense, is the generation of action items. These action items, which may or may not specify bugzilla issue(s) being opened.

<BassamT> scribe, bassamt

<BassamT> scribe: bassamt

<Paul> ... When action items are opened without mentioning bugzilla, it is up to the tasked person to decide if they will open bugzilla issues.

<Paul> ...This WG will review as a body the action items when they are reported as completed before they are closed.

<Paul> ...suggest we explicitly agree on tools use. Hence the suggestion above.

MSM: is sml:ref="false" meaningful.

Bassam: Yes, in cases where a reference element shows up in an XML document that has not been validated.

Ginny: What about invalid values for sml:ref, such as "10" "foo" etc.

<scribe> ACTION: Bassam to change section 3.2.1.4 to handle cases for invalid values of sml:ref="" including "false". [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action16]

<trackbot-ng> Created ACTION-18 - Change section 3.2.1.4 to handle cases for invalid values of sml:ref=\"\" including \"false\". [on Bassam Tabbara - due 2007-06-19].

<scribe> ACTION: BassamT to change the spec to support all valid values for sml:ref="true", including sml:ref="1", etc. [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action17]

<trackbot-ng> Sorry, couldn't find user - BassamT

<scribe> ACTION: Bassam to change the spec to support all valid values for sml:ref="true", including sml:ref="1", etc. [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action18]

<trackbot-ng> Created ACTION-19 - Change the spec to support all valid values for sml:ref=\"true\", including sml:ref=\"1\", etc. [on Bassam Tabbara - due 2007-06-19].

<Bassam> Ginni: can the input node set for a reference R be matched to the output node set?

<Bassam> Sandy: does deref() validate targetRequired cosntraint.

<Bassam> Bassam: No.

<Bassam> Ginny: will think about whether there are scenarios for correlation of input and output reference node sets.

<Bassam> MSM: why does deref() return one target element only? This seems limiting.

<Bassam> MSM: will take this offline and raise a bugzilla issue if he deems it worthy.

<Bassam> ACTION: Sant to add bugzilla item on section 3.2.1.4 WRT to supporting the case when deref() operates on a reference that resovles to multiple target elements. [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action19]

<trackbot-ng> Sorry, couldn't find user - Sant

<Bassam> ACTION: Sandy to add bugzilla item on section 3.2.1.4 WRT to supporting the case when deref() operates on a reference that resovles to multiple target elements. [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action20]

<trackbot-ng> Created ACTION-20 - Add bugzilla item on section 3.2.1.4 WRT to supporting the case when deref() operates on a reference that resovles to multiple target elements. [on Sandy Gao - due 2007-06-19].

<Bassam> ACTION: Bassam to normalize usage of the term "reference element" for the parent and "reference scheme" as the child element. [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action21]

<trackbot-ng> Created ACTION-21 - Normalize usage of the term \"reference element\" for the parent and \"reference scheme\" as the child element. [on Bassam Tabbara - due 2007-06-19].

<Bassam> ACTION: Bassam to remove last setence of section 3.3.1 second bullet "... and MUST attempt to resolve ..." [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action22]

<trackbot-ng> Created ACTION-22 - Remove last phrase of section 3.3.1 second bullet \"... and MUST attempt to resolve ...\" [on Bassam Tabbara - due 2007-06-19].

<Bassam> ACTION: Bassam to change "SomeValidUri" sample in first example of section 3.3.1 to something more real. [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action23]

<trackbot-ng> Created ACTION-23 - Change \"SomeValidUri\" sample in first example of section 3.3.1 to something more real. [on Bassam Tabbara - due 2007-06-19].

<KirkD> KirkD is Kirk

<Bassam> ACTION: Bassam to change all example URI of the form "/foo/bar" to absolute URIs of the form "http://example.org/foo/bar". [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action24]

<trackbot-ng> Created ACTION-24 - Change all example URI of the form \"/foo/bar\" to absolute URIs of the form \"http://example.org/foo/bar\". [on Bassam Tabbara - due 2007-06-19].

<plh> URI = scheme ":" hier-part [ "?" query ] [ "#" fragment ]

<plh> URI-reference = URI / relative-ref

<Bassam> ACTION: Bassam to change all usages of URI to URI-reference which includes both absolute and relative forms of URIs. [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action25]

<trackbot-ng> Created ACTION-25 - Change all usages of URI to URI-reference which includes both absolute and relative forms of URIs. [on Bassam Tabbara - due 2007-06-19].

<Bassam> action- 24

<plh> XPointer Registry

<Bassam> ACTION: MSM to open a bugzilla item on whether other fragment identifiers should be supported in the URI reference scheme. [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action26]

<trackbot-ng> Sorry, couldn't find user - MSM

<Bassam> ACTION: Michael to open a bugzilla item on whether other fragment identifiers should be supported in the URI reference scheme. [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action27]

<trackbot-ng> Created ACTION-26 - Open a bugzilla item on whether other fragment identifiers should be supported in the URI reference scheme. [on Michael Sperberg-McQueen - due 2007-06-19].

<Bassam> ACTION: Michael to open a bugzilla item on whether XPointer should be the only fragment identifier scheme supported in the URI reference scheme [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action28]

<trackbot-ng> Created ACTION-27 - Open a bugzilla item on whether XPointer should be the only fragment identifier scheme supported in the URI reference scheme [on Michael Sperberg-McQueen - due 2007-06-19].

<Bassam> ACTION: Michael to think about how different MIME types relate to SML. [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action29]

<trackbot-ng> Created ACTION-28 - Think about how different MIME types relate to SML. [on Michael Sperberg-McQueen - due 2007-06-19].

<Bassam> ACTION: Bassam to fix EPR example in 3.3.2 to not include reference params that identify a resource. [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action30]

<trackbot-ng> Created ACTION-29 - Fix EPR example in 3.3.2 to not include reference params that identify a resource. [on Bassam Tabbara - due 2007-06-19].

<Bassam> ACTION: Philippe to open a bugzilla item on whether EPR reference scheme should be removed. [recorded in http://www.w3.org/2007/06/12-sml-minutes.html#action31]

<trackbot-ng> Created ACTION-30 - Open a bugzilla item on whether EPR reference scheme should be removed. [on Philippe Le HÃ©garet - due 2007-06-19].

- DRAFT -

SML Face-to-Face

12 Jun 2007

Attendees

Contents