SML WG Teleconf 2007-09-06

 

 

<johnarwe> http://www.w3.org/XML/Group/SML

<scribe> Scribe: Jordan Boucher

<scribe> ScribeNick: Jordan

<johnarwe> updates to agenda scribe list

<johnarwe> 2007-08-23 Eckert, Zulah

<johnarwe> 2007-08-28 Lynn, James

<johnarwe> 2007-08-28 Gao, Sandy

<pratul> Agenda is at http://lists.w3.org/Archives/Public/public-sml/2007Sep/0009.html

<johnarwe> 2007-08-29 Valentina Popescu

<johnarwe> 2007-08-29 Wilson, Kirk

<johnarwe> 2007-08-30 Lipton, Paul

last meeting minutes

resolution: minutes approved
... skip over due actions this week since Tracker website is unresponsive

item 5 - Editorial continuity

resolution: Jim to lead editors through process of getting draft to webmaster while Ginny is out

<johnarwe> http://lists.w3.org/Archives/Public/public-sml/2007Sep/0007.html

item 6 - Continue references discussion

Sandy: summarized issues
... need agreement on sections 2, 3 and 6

<MSM> [Pratul, can you expound on why you believe the identity checking problem is hard? Where does the difficulty lie?]

Ginny: shares Pratul's concern about section 2

<pratul> Here's the URL for my email http://lists.w3.org/Archives/Public/public-sml/2007Sep/0019.html

discussion about deref()

<pratul> MSM, the identity check is possible to implement using the notion of a document identity and the XPath from the document root to the element

<pratul> However, there was concern in the private WG that this will be difficult to implement in distributed scenarios where the model docs are distributed

<johnarwe> Ginny suggested de-coupling def of deref() from consumers. Kumar notes "consumer" was recently removed from SML spec.

<johnarwe> e.g. ref proposal 2.1 is "strict", 2.2 is "lax". validating consumers require "strict" behavior. this separates the concerns.

Ginny mentioned need to update conformance section based on this decision

resolution: de-couple def of deref(), as suggested by Ginny

<pratul> resolution is to decouple the defn of deref() from the notion of a consumer

<johnarwe> MSM suggests de-coupling ref validity checking (eg mult ref schemes in 1 ref must pt to same doc+element)

<johnarwe> ... from definition of deref()

<johnarwe> Pratul notes current content of deref() (x.1.2.4) only has a vague implication that the ref's validity gets checked

<johnarwe> section is structured oddly...validation implication is really implied by pre-conditions.

<johnarwe> Sandy suggests ref proposal 2.1 be used for what validators must do, and 2.2 be used for what deref() must do.

<johnarwe> awkward section is http://www.w3.org/TR/2007/WD-sml-20070806/#deref_28_29_XPath_Extension_Function

<johnarwe> MSM would like to say "an impl of deref() that actually checks all the schemes is just an impl of 2.2"

John: continue ref proposal via email

Updated Scribe List

200y-mm-dd        Vijay Tewari 
2007-07-26        Kumar, Pandit 
2007-08-09        Waschke, Marvin
2007-08-16        Smith, Virginia           9/13 through 9/27
2007-08-23        Eckert, Zulah 
2007-08-28        Lynn, James 
2007-08-28        Gao, Sandy 
2007-08-29        Popescu, Valentina
2007-08-29        Wilson, Kirk 
2007-08-30        Lipton, Paul 
2007-09-06        Boucher, Jordan