[wsia] [minutes] 7/12 (resend)

[Graeme Riddell <griddell@bowstreet.com>]

OASIS Web Services for Interactive Applications
 Technical Committee

12 July 2002
teleconference 12:00-1:00pm EST

Minutes



Attendance

	William Cox		BEA
	Graeme Riddell		Bowstreet		y
	Greg Giles		Cisco
	Srinivas Vadhri		Commerce One		
	Kirk Wilson		Computer Associates	y
	Sean Fitts		Crossweave
	Tim Jones		Crossweave		y
	Peter Quintas		Divine
	Sandra Swearingen	DoD
	Monica Martin		Drake Certivo		y
	Alan Kropp		Epicentric
	Patel Ashish		France Telecom
	Aditi Karandikar		France Telecom
	Royston Sellman	HP
	Eric van Lyndegraf	Kinzan			y
	Charles Wiecha		IBM (chair)		y
	Dan Gisolfi		IBM
	Ravi Konuru		IBM
	Rich Thompson		IBM			y
	Shankar Ramaswamy	IBM
	T.V.Raman		IBM
	Rex Brooks		individual		y
	John Kneiling		inidividual		(removing from the
list)
	Kevin Brinkley		Intel
	Michael Mahan		Nokia
	Martin Chapman		Oracle
	Terry Cline		Peregrine Systems
	Sasha Aicken		Plumtree
	Stefan Beck		SAP
	Howard Melman		Silverstream
	Jeffrey C. Broberg	Silverstream
	Suresh Damodaran	Sterling Commerce
	Eilon Reshef		WebCollage
	Gil Tayar		WebCollage

Agenda

1. Roll call and minutes
2. Update on joint interface spec and WSIA goals for the next draft
3. Plans for preparing a draft Customization spec in parallel with the joint
interfaces.



Meeting

12:00	roll call + minutes.

To fulfill the decision made prior to the F2F of staggering these calls with
the WSRP calls we'll meet again Friday of next week (19th) and then the
following Thursday (25th) and then every other Thursday. Rex is now the
webmaster and is bringing the site up to date. Subcommittee minutes should
be forwarded to the list so they can be posted too.


12:10	Status of joint interfaces and WSIA-specific goals

0.21 of the joint spec is coming out today. Also starting the process of
putting a User Guide together (strawman was developed at F2F). Smaller group
also looking at an object model and extensibility model that might lead to
some refactoring and layering - Policy and Design issues came up at the F2F.


Day 1 of WSRP suggested that the interface was made too complex by trying to
do stateful/stateless, persistence, user state, etc. Rich leading effort to
back up and see if we can identify separate objects in the interface to
simplify the functionality. This might allow easier extension and selection
and pluggability of the functionality.

This might reintroduce choice of a vocabulary of operation names and
signatures that are a little more specific to WSRP and WSIA. (We've had the
idea that there would be a joint spec that accommodated WSRP with a single
interface and set of operations. That interface is broader than expected, so
we're starting to feel (and investigate) that there may actually be separate
interfaces that work alongside each other.)

Thoughts on this?
- So long as we figure out how to piggyback WSIA on top of the base protocol
it should be ok? But will we then have a joint spec? The joint spec might
then be the  Embedded case.

- Can we not build on top of WSRP? Two concerns are then a) operation naming
would likely be too portlet-specific, and (b) danger that a portal-specific
API would not be able to support non-portal consumers.

- General rule should be same that the same function should not be allowed
in both the WSRP and WSIA interfaces?

We should be able to achieve a joint spec at the component/lifecycle level.
The entity stuff might become specific to WSRP. Aim to have a joint spec at
the lowest level that both WSRP and WSIA can build on. But we're probably in
trouble if the bottom of the stack is not common - we'd end up with 2 specs.


12:35	Customization (Tim)

Latest thinking is to piggyback cutomizable properties on top of the joint
protocol and reuse joint interfaces. Specifically starting to work on
initializing model instance values on the producer.

Also given the current factoring effort, we probably have the opportunity to
carry forward a property interface.

Point of clarification - need to emphasize that the property interface does
not extend the opaque state of the producer but it is a view of the opaque
state (of the producer properties).

The other issue is how and where those properties are described. Tim & Ravi
leading this effort.

A draft is expected by September F2F in conjunction with 1.0 of the joint
spec. The property description stuff especially should be orthogonal to
joint interfaces spec. 

A: Tim - Aim to have an outline of the draft document by this Wednesday's
Customization conf call.


12:40	other stuff

Next F2F in front of WSRP on Sept 9th-12th, hosted by IBM in
Munich/Stuttgart.
Next conf call Friday 19th.


12:45	adjourn.