Minutes of 2008-10-30 telcon

[Anish Karmarkar <Anish.Karmarkar@oracle.com>]

anish: Scribe: anish

anish: Agenda:
1. Roll Call

2. Scribe Assignment

3. Agenda Bashing

4. Approval of minutes of 10/16/2008 conf-call
http://lists.oasis-open.org/archives/opencsa-liaison/200810/msg00010.html

5. Issues

6. OpenCSA Project Management
Review of testing plans

Version Control System recommendation status/update

Any other blocker issues?

7. Next conf-call?

8. AOB


Topic: Respository

anish: Mike: Few things to talk about: one of them is version control system

anish: SVN seems to be consensus

anish: OASIS is working on it

anish: Jeff: when do we need this by?

anish: Mike: we just started going. another couple of weeks before we 
have any artifacts. Doable if we get something by end of november

anish: Dave: definitely need it by end of year. Hard to say if we'll 
absolutely need it before TG

anish: Dave: be willing to be guinea pigs if OASIS wants to provide us 
something

anish: Will need an SVN by begining of december

Topic: Approval of minutes

anish: Mike moves to approve minutes of 10/16/2008 located at 
http://lists.oasis-open.org/archives/opencsa-liaison/200810/msg00010.html. 
Dave Seconds

anish: no objections. Minutes are approved

anish: Topic: status reports

anish: Assembly - 9 open issues left. 2 new ones being created

anish: ... Have a complete pass of 2119 statements and test tags

anish: ... out for review

anish: Policy - 9 or 10 issues open. Resolution text in new draft, need 
approval. Just behind Assembly. On track

anish: Java - quite behind. f2f coming up, so hope to make some progress 
during that time. 50 new issues. F2F has 3 new days.

anish: ... a lot of those issues are from spec review and are minor

anish: ... The chairs will have a better feel for where things are after 
the f2f

anish: BPEL - 2 new issues open. One new issue filed. Editors have 
provided a draft for RFC 2119 and is currently under review.

anish: Bindings - 17 open issues and 3 new one opened today. New 
contribution for binding.http. If that is accepted then there will be 
more issues, but unclear whether such a new binding would be on the same 
schedule

anish: C/C++ - Bryan not on the call

Topic: Schedule and exit criteria

anish: Jeff: how are we doing on the schedule?

anish: Dave: i think we're behind

anish: Jeff: do we need more data?

anish: Dave: after the java f2f might be a good time to talk about 
schedule again

anish: MikeE: we also need 2 impls for every spec. We haven't asked the 
Q of where those implementations would come from

anish: Jeff: consider that to be part of testing group

anish: MikeE: tuscany may be able to do that but depends on how we cut 
the cloth and how rigorous we want to be

anish: Anish: do we need two implementations for all features or only 
mandatory features

anish: Jeff: we need to two impls. for all the features

anish: ... otherwise you have really said anything

anish: s/have/haven't/

anish: MikeE: all impls should pass the test suite, but don't think 
we'll have a test suite to test every feature

anish: Dave: does OASIS have any guidance wrt this?

anish: Jeff: we have to define it

anish: Jeff: have to convince two-thirds of the SC

anish: Mike: problem is resources

anish: Dave: what if the test committee test only the mandatory aspects 
of the spec?

anish: Jeff: will have to think about it. That should be the minimum bar

anish: Jeff: we don't want this to be an invitation for someone to 
profile it

anish: Jeff: also a Q of Committee spec v. OASIS standard

anish: ... for example, we may approve CS with minimum bar and require 
more for OS

anish: Dave: any time issues between CS and OS

anish: Jeff: no maximum, minimum is 7 days

anish: Jeff: the idea was that OS is more mature

anish: Jeff: we could do a staged roll out

anish: ... throwing out ideas

anish: MikeE: we should aim to pass the giggle test, avoid profiles by 
creating a test that contains a lot of features. But writing a test for 
every feature would be hard

anish: Martin: could just lay out the mandatory part

anish: MikeE: not sure we can create tests for *all* mandatory parts

anish: Jeff: this needs more thought

anish: Dave: need to think about subsetting idea more

anish: Jeff: we can spend a lot of time discussing in abstract, but we 
need to starting doing things and get an idea of what the gap is

anish: MikeE: wrt Java, there are lot of things, say with threading, 
that will result in lot of tests

anish: Dave: yes, we could potentially write 100s of tests for Java

anish: Jeff: like the idea of including a SCDL that has all the 
features, so implementations have to atleast support all the syntactic parts

anish: Jeff: this conversation will have to continue

anish: Topic: next meeting

anish: Mike: ok having a meeting next week, but not the week after (java 
f2f)

anish: Jeff: won't be available next week, but sanjay may be

anish: Jeff: any point in having a call till after the java f2f?

anish: Dave: unless there is something about SVN, not much to talk about

anish: Jeff: next call on 20th nov, unless something comes up about SVN

anish: Meeting adjourned