Groups - OpenCSA LSC Regular Conf-Call modified

DRAFT agenda Open SCA Liasion SubCommittee

07 Jan 2008

Dial-in: 877-633-8727/722572
Web-conf: http://webconf.soaphub.org/conf/room/sca-lsc

Chair and notes: Jeff

1.	Roll

Chairs:
Sanjay Patil, Jeff Mischkinsky

Assembly: Mike Edwards, Martin Chapman

Policy:
Dave Booz, Ashok Malhotra

Bindings: Simon Holdsworthy (absent)

BPEL:
Sanjay Patil, Anish Karmarkar

C/C++:
Bryan Aupperle

J:
Mike Rowley, Henning Blohm (absent)

2.	Agenda Bashing

no changes

3.	Approval of Dec 10th meeting Minutes

RAW IRC log: http://www.oasis-open.org/apps/org/workgroup/
opencsa-liaison/email/archives/200712/msg00024.html

(Do we want minute taker to provide the "official" summary?

APPROVE minutes dec 10 - mike/bryan - UNAN

4.	Review Action Items

no outstanding AIs

5.	Issue Management

Current issues in JIRA:
http://www.osoa.org/jira/secure/IssueNavigator.jspa?reset=true&mode=hide&sorter/order=DESC&sorter/
field=priority&resolution=-1&pid=10040&component=-1

5a. Any proposed new issues? - I see no outstanding emails, so
I assume not.

no new issues

6.	Issue Discussion

6a. liasion-1 -  Define Conformance Targets - http://www.osoa.org/
jira/browse/LIAISON-1

There is a concrete proposal in jira. More discussion
needed? Can we come to a resolution "soon"?

Discussion about specs being "conformance targets". Assembly is the
"main" spec in that it defines the topmost level of SCA conformance
and puts requirements on other specs. There was general agreement on
this being the way to move forward.
  Mike R there 2 kinds of conformance targets: SCA runtime, any SCA
artifacts. SCA binding, as an artifiact, might have conformance
statements. SCA binding, as a part of the SCA runtime, might have
conformance statements about its behavior.
  MikeE notes that XSD conformance (when consuming a doc) can be
handled by a generic sort of statement about checking for XSD
conformance. Additional structural/semantic requirements on docs,
have to be separately specified.
MikeR pushes back on the usefulness of defining specific error points
since we don't know when and how errors will be caught.
Anish asks how to define errors independent of languages.

Modify point 2.
Point 3 kinds of errors need to be defined relative to the
appropriate runtime behavior in question - e.g. deployment, a
language C&I, etc.

Dave Booz is concerned that the large policy algorithm places lots of
different requirements on different artifacts, processors, etc., and
that he is not sure we have the means to express those conformance
requirements yet.

Sanjay: Identifying the points of raising errors and standardizing
the errors, if needed, is a matter of completeness of the
specifications. Phrasing conformance targets is a separate matter.

Michael Rowley: Conformance targets can be roughly categorized into
1) document artifacts (or constructs within them) that can be checked
statically.
2) SCA runtimes, which we may require to exhibit certain behaviors.

Michael Rowley: 2. For each appearance of MUST or MUST NOT, specify
the construct or runtime behavior that is being constrained.

Folks should continue on email a/o JIRA with some attempts to refine
the proposal in liasion-1

6b. liasion-2 - Usage of RFC 2119 terms with lower-case - http://
www.osoa.org/jira/browse/LIAISON-2

more to be said? Do we just need to take a vote, if not at
this meeting at the next one?

We didn't get to discuss this.

7.  Schedule of next meetings
next meeting is next monday 14 jan 2008  same time (0800 PST) and
same call-in

8. AOB
  none

--
Jeff Mischkinsky
		
Director, Oracle Fusion Middleware and Web Services Standards	+1(650)
506-1975
Consulting Member Technical Staff
			500 Oracle Parkway, M/
S 4OP9
Oracle								Redwood
Shores, CA 94065