Minutes meeting Jan 19

["Jacques R. Durand" <JDurand@us.fujitsu.com>]

For review:

 

-jacques

Test Assertions Guidelines (TAG) TC meeting

Event Description:
------------------

Tuesday 2pm (California time), 19 January 2010

Host confcall: Fujitsu
US Toll Free: 877-995-3314
US Toll/International: 210-339-1806
Passcode: 9589308

Agenda: 
-------

1. Administrative:
- approval min Jan 5

2. Review of the three candidates CDs:
- Model
- Mark-up
- Guidelines

3. Decision on initiating a ballot:
- any last updates that need be done before initating a ballot?
- ballots CD + PR separate?


Participants:
------------

Stephen Green, (Document Engineering Services)

Jacques Durand, (Fujitsu)

Tim Boland (NIST)

Paul Rank (Sun)

Kevin Looney  (Sun Microsystems)


Action Items:
-------------

AI1-Jan19: [All] make sure to review the 3 candidate docs and comments on mail list.

AI1-Dec08: [Dennis] check availability of JIRA (official OASIS debug tracking).

AI2-Dec08: [jacques] check with Sun if Dimitri still available for helping 
on tech writing/review.



Minutes: 
-------


'''1. Administrative '''

- Postpone the approval of minutes Jan 5th.


'''2. Review of the three candidates CDs '''

- Comments from Dennis, Kevin, Jacques.
- most significant comment from Kevin [5]: presence of the Report element in the TA model, 
that was questioned: too close to implementation, could be added to any "representation" 
or XML as these could extend the model. 
Majority was in favor to remove from model and keep it in the mark-up.
- SG: This change alone will require some time, at least a week part time. Lots of references.
- discussed other "design" comments from Jacques [11](complex derivedSourceItem) and[19]
(should conf clause be extended). [11] needs more thoughts - but is only a simplification issue. 
- discussion on [19] whether a TA instance - using a formal representation/markup, 
or even just using an informal notation - could be said conforming to the Ta MODEL	
(meaning model conf clause would then cover TA instances as a conf target, 
in addition to a representations/markups as conf target). 
Conclusion is that is probably not necessary: conf to the model is only meaningful 
for formal representations.

- discussion on issue that although some TA parts are mandatory (target, predicate..)
their cardinality in the model allow these elements to have no internal component at all (0..1).
That seems inconsistent...
- JD: For Predicate we can madate that "content" be always there. For Target, we could either:
(a) mandate that either Content or Type always be there (no two of them absent) via a rule
that representations will have to enforce or require, (b) require at least content be non empty.
- SG: favor (b).
- JD: but some TAs may just want to specify a "type" as they apply to a category.

- SG: there is still an issue about TA set and doc Header element.
- JD: the most pressing area of teh design that needs be secured first and cleaned-up is the 
TA itself: higher level constructs are less urgent (and fewer readers will be itnerested in them)

- most others comments were editorial.


'''3. Decision on initiating a ballot '''

- enough internal comments on 3 specs were made to justify working on these 
until next regular meeting, 2 weeks from now. 
- planned best scenario: 
(a) resolve issues by email over next 2 weeks.
(b)next meeting we review, possibly agree on final modest editorial changes, 
get a consensus that the specs are OK to vote about.
(c) 2 or 3 days later, Stephen posts final candidates.
(d) I initiate an electronic ballot for CD, and for PR.
(e) end of ballot (1 week) I submit PR request to OASIS early February.
- We can also, in case Editor is short of time, extend (c) of about 1 more week.
- All 3 docs go for PR together.