OASIS Web Services for Remote Portals (WSRP) TC
WSRP Conference Call - August 8, 2002
Roll Call
Alan Kropp, Epicentric
Rich Thompson, IBM
Carsten Leue, IBM
Thomas Schaeck, IBM
Charlie Weicha, IBM
Lothar Merk, IBM
Mark Cassidy, Netegrity
Mike Freedman, Oracle
Sasha Aickin, Plumtree
Yossi Tamari, SAP
Stephen White, SeeBeyond
Ugo Corda, SeeBeyond
Peter J Quintas, Divine
Andre Cramer, Citrix
Tim Jones, Crossstreet
Adrian Fletcher, BEA
Susan Levine, PeopleSoft
Rajesh Tiwari, PeopleSoft
Michael Hillerman, PeopleSoft
Agenda
Two main agenda items:
But first we should discuss organization for F2F.
Discussion
Lothar: agenda and hotel info posted on website, please register ASAP, Monday wsia meeting, wsrp members also invited.
Other: issues related to property interfaces, property description format, relevant to wsrp as well as just wsia
Thomas: so we just need to send a note to Lothar about attendance.
Thomas: Next item, report on spec program, Carson and Rich.
Rich: okay v .3 of spec has moved into oasis template for upcoming standards. I'd done rearrangement, variety of things that were design oriented, now all in section 3, not well ordered yet. Various interface sections attempt to go in order or how the service is used. description, register, deal with entities, get markup, transparent state.
large set of changes flowing in, next week, new version on email list.
email is good way to initiate discussions.
Thomas: okay how do we proceed from here, so we had aimed for this week to have a baseline that we can use to give people opp to implement and experiment with what's currently in the spec. what's a good time to actually do this? a little too early is Friday, what' the next point in time?
Other: I'd suggest a mid-next-week timeframe, there's some sections like security that need to be brought in, that would give the opportunity to be brought in.
My anticipation is the main mechanism of the protocol is fairly firm. We're tweaking at this point instead of rewriting. Esp. since we made the decision earlier this week about sessions.
Thomas: I'd agree w/current draft and ideas on the sessions, we might be able to give a prototype of a reference implementation.
Mike F: This is a good time to talk about the process of getting that work started. At the f2f we had a number of people who volunteered to contribute to that. Let's get more detailed on how that's going to work.
Thomas: I think the one other thing that comes to mind is we'd kicked off a discussion about the underlying object model, a couple versions of the diagram of how to represent it, we should review that and if we agree t hat it represents our concepts, does that have impact on some of the signatures of the operations. Would that be reasonable to do on Tuesday �lan do you think?
Alan: yeah, don't see why not. I'd be more inclined to flush out discussions we've had,
We may actually hit the protocol, how session handles are handled...
Thomas: okay we'll add that to the agenda.
okay, summarize: 1st when to we have baseline to recommend to try implementing? so we said around middle of next week.
Other: so you think discussion of underlying object model could have impact on protocol?
Other: could come back to whether we can handle ??? or ??? could have impact on operation signatures.
Mike F: I think the details will be fluid for some time, at least until the f2f ...
I think people who are doing implementations are willing to track those kinds of changes, that's fine.
baseline - it's not going to be stable, should wait until after f2f, if they want stability.
Thomas: so I propose we pick next week, let's pick friday next week, we'll have a version of the spec we recommend for starting with understanding there will be changes in the spec.
Mike F: do we also want to bring committee work on implementation?
Thomas: we have to clarify issues about code sharing and all that, on that side, everybody who wants to participate should check with your attorneys of the individual companies, then also to incorporate some kind of open source project, also companies figure out the policy of participating.
Mike F: one of the things it might be useful to clarify is what are the work products, maybe you could write that up? at the f2f we talked there's a sample impl. for test suite, performance suite, but we've been saying it would be interesting to provide impl of jsr168, so we need to be clear.
Thomas: i was saying we produce software on which you can put wsrp services, one way is to plug in the jsr reference impl when available, that has to be ultimate goal, would you agree?
Mike F: that would be a great sample or ref impl, yes.
Thomas: ?? making a point have to decide about language for this implementation, for this reason, portability, etc. it seems quite ??? use java.
Mike F: yeah, I think that's also very natural, running on tomcat, supporting jsr168.
Thomas: next thing to agree on is open source project, if we want to base on tomcat, to ??? actually
Mike F: how would you propose we go about making these decisions, should the whole group do it, or talk to people who indicated interest at the f2f, over the mail?
Thomas: I think we should setup a separate call for this to have a more in depth discussion about this topic, let me sched for next week, and I can put some preparation and next week we can discuss schedule for impl and can discuss how it should look, etc.2
Thomas: okay next agenda item for today, discuss issues and questions regarding spec that group as a whole might have.
Mike F: a high level question: 2 wks ago we talked about spec being expert group draft or needing to be presented to outside world, talked about producing a toc for the latter for an outline of the info that we'll need... did we make any progress?
Alan: no i've not made any progress
Other: the other half of that was the expectation was that the intro material would come out of the white paper that thomas has circulated a first draft of (which is on email and archive)
Thomas: have got feedback, i want to circulate update of that. It should be suitable for an introduction.
Mike F: do you think that we actually need that level of detail from a wsrp perspective? why do we care about simple consumers for example? might we not have a structure that talks about the prototypical producer and allude to later, not highlight as much, the other kind of producers that you might have to deal with. or is the intent not to end up with a wsrp spec but a wsia spec and we tell folks to read wsia spec and figure out common use cases for wsrp.
Carsten: from my impression would agree that the consumer would not make sense in wsrp spec but producer certainly does. in the beginning there would be portal service serving...??? various producers not rely on any portal service, would be important part of spec.
Mike F: wasn't asking that it be excluded, but talking about how it might be structured in the spec. to put words in joe's mouth, joe, ?? pick this thing up and read about, I've been studying portals, want to see what spec is all about, wouldn't it be clearer if the spec allows me to quickly get in and understand from the perspective that I already know what's going on, then goes into broader generalities, as opposed to just laying it out there as a uniform set of capabilities and then I have to dig around and map that back to my understanding.
Other: didn't we talk about implementation guide?
Other: but joe's a consumer of this, not an implementer, he needs to understand whether it meets his requirements in terms of representing portlet.
Other: but this is a joint spec, this is a formatting issue. we can package it to reflect the nature of the joint work...
Other: that's what I'm saying here...
Other: from a wsrp perspective do we want to have a different formatted spec or first couple of chapters of the spec than the broad generic wsia version of it.
white paper is to provide general overview, which is why wsia has section with simple use cases, going to complex use cases.
Mike F: I'll read the white paper and make comments, I'm not sure it's pertinent to end reader community as it was for us to do in the f2f because we were struggling with decisions on how to represent things in the protocol, and we needed a common basis. when you turn around and say someone who's reading this is familiar with portlets, by ibm, bea,oracle, plumtree, whatnot, should we have something that makes it easy for them to get in and understand the spec w/o having to dive thru use cases that don't apply as strongly to portlets.
Thomas: at very beginning comes motivation of wsrp that starts by saying... there are portals, local portlets, can't just plug and play components on remote server...
essentially my comment is take one of your customers Thomas, and would you expect them to pick up our draft, and very easily understand how the 2... understand the spec
Thomas: I would say they need the white paper to understand the spec
Mike F: I think we should figure out a form of the spec to allow our readers to pick the thing up. some of the extra burden in wsrp in terms of writing incur over this generic spec.
??white paper cover that from a high level, organizations that want to provide connection do their own doc?
Mike F: but in the end all the customers are going to have to look at the spec an understand it because their bosses tell them to, but most folks will be able to ignore it because container hides the details, but for those folks, how do they get into it, and I think that's the responsibility of this group.
Thomas: yes, wsrp should provide all the doc, white paper, and implementation guide, spec..
or maybe a reader's guide, another thing that has the overview, sections to concentrate on, I'm not sure.
other option, make spec understandable if you're read the white paper.
Thomas: okay, let's get back to Alan's toc for a moment.. when we get the baseline, can we ????
Alan: okay, I'll want to collaborate with a few folks on that... I guess I can read the white paper myself, do that and take another pass at the spec and abstract out what I think would be useful toc line items, I'll have that next friday.
Thomas: okay, let's look at coming updates of the spec, give feedback so we can make progress until next week.
Thomas: Does anyone else have any other questions or comments about the spec?
Meeting adjourned.