Logistics
Roll-call
Other Business
Business In Order
Action Items & Motions
References
Meeting Date |
03/10/2003 |
Meeting Time |
10:00 pm CDT |
Location |
Dial-in Number: 888-742-8686 /
303-928-2600 (Outside of the United States/Canada) |
Duration |
1 Hour |
Chair |
Darran Rolls |
Recording Secretary |
Gavenraj Sodhi |
Agenda |
As published as of March 10th, 2003 |
Present
(GS) |
Gavenraj Sodhi, Business Layers
|
(ML) |
Matthias Leibmann, Microsoft |
(JF) |
Jesus Fernandez, Computer Associates |
(HL) |
Hal Lockhart, BEA Systems |
(DR) |
Darran Rolls, Waveset |
(JB) |
Jeff Bohren, OpenNetworks |
(DC) |
Doron Cohen, BMC |
(RE) |
Rami Elron, BMC |
(PM) |
Paul Matson, Entrust |
(GW) |
Gerry Woods, IBM |
None |
|
1 |
Vote to minutes of committee meeting 02/24/2003 |
DR |
Minutes available at [1] – Approved |
Proposed Agenda
===============
1 - 10:05 Order and role-call
-----------------------------
Gavenraj to take roll-call and summary
minutes please.
2 - 10:07 Accept meeting minutes
--------------------------
From 3/3/2003 meeting, available at [1] -
ACCEPTED (8:09 a.m.)
3 - 10:40 ResultCodes
---------------------------------
Recap on the model. Agree current list of resultCodes:
Success
Failure
Pending
malformedRequest
UnsupportedOpeation
UnsupportedIdentifierType
NoSuchIdentifier
Doron:
Posted proposal to separate result code and reason code… to have definite result code to say if
operation has success or failure.
Result code should be extensible so each vendor can add their own codes.
Comment:
A two level code should be good and some way should express unsupported
operations.
Darran:
Is anyone opposed to two-levels of code?
Comment: Reason code should not be extensible.
Result Code – Reason Code à to be able to look for more help.
Reason code should be extensible which should
be done in operational attributes.
Comment: Placeholder should exist for Reason Code
MOTION and Action Item: Result code and reason code are to be defined within enumerated
list within
-
ACCEPTED
Action Item:
Have XML Lang to do error handling.
4 - 10:45 Singleton Requests
---------------------------
As per the discussion thread in [3] discuss
and conclude on making
SPMLReeust support ExecutionType and hence
make SPMLBatchRequest optional.
Comment:
Expression of what is supported comes out as a WSDL.
Comment:
Would there be separate Core Types?
- Not necessarily. There could
be one WSDL files that describes all SPML Services.
MOTION:
Generically, SPML v1 should support single term request and batch - ACCEPTED
MOTION:
Remove search request and schema request from the batch request –
ACCEPTED
Table:
Putting into WSDL definition (Operations Query, Agenda Item #5) – ISSUE
Item which Gerry will Champion
5 - 10:55 Operations Query (Tabled)
---------------------------
As per the discussion thread in [3] discuss
and conclude on providing a WSDL model for client query of supported operations
in 1.0.
6 - 10:10 Gerry's WSDL Proposal
-------------------------------
As discussed on last weeks call, Gerry
prepared a detailed outline of his alternative proposal for a more WSDL/Schema
focused model for SPML (available at [2]).
Having reviewed this proposal, the following are the questions before
the committee:
- General questions for Gerry on his proposal
- Does this proposal offer a better model for
SPML
- If so what is the impact of making this
change and what will be affected
- If not, what can we learn from this
proposal
- Next steps (if any)
Gerry’s Comment: Allow for the schema to be more expressive
Comment:
Problems of making this more provisioning centric. Don’t know how filters will be supported.
Comment:
For filtered searches, making the syntax a multi-valued would help
resolve any problems. Maybe doable with
XML Schema.
Comment:
Would Support attribute multi-valued pair.
Comment:
Jeff to put out filtered search in his proposal
DR:
Make decision around, can we move ourselves forward or do we need to
look more closely at where we are.
Comment:
We have fundamentally a rework or our discussion in Redmond.
Comment:
Could deliver Provisioning Semantics with current extended request
definition. Did talk about pushing out
a SPML Schema for version 1 or further.
GW:
You would probably end up with two specs.
Comment:
What would we need to add to Extended Request model?
DR:
Show how we may use Expressive Schema within model.
7 - 11:00 Motion to Adjourn
---------------------------
To reconvene 3/17/2003.
[1] http://www.oasis-open.org/committees/provision/minutes/pstc-minutes-03032003.html
[2] http://lists.oasis-open.org/archives/provision/200303/msg00022.html
[3] http://lists.oasis-open.org/archives/provision/200303/msg00015.html