OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

wsbpel-implement message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]

Subject: Re: [wsbpel-implement] Reminder - call Monday Nov. 11, 1pm Eastern

    Hello all.  Pleased to hear you're moving into the running-code stage.   Let me mention a few things that may be relevant to the "testbench" trial.  This responds to Diane's request for comment, below.  Feel free to ping back with questions if this is unclear or doesn't seem to meet your needs.

    The bottom line rule for OASIS is that all TC work is publicly readable, so any testbed artifacts, test specs, result grids, etc., that are produced will need to be posted to the lists, where FYI they will be eventually visible to the public archives.
    However, while we can't do "secret", we can do "civilized".  It's a perfectly reasonable technical goal to conduct civilized tests of alpha implementations without a lot of fanfare, so that members can confidently bring experimental code to the table without being embarrassed.  But in OASIS we don't have NDAs or the like, and we don't have the authority to mandate confidentiality covenants or exclude TC members from TC activities -- so we have to use social mores instead of lawyers to accomplish this.  But this is still doable.
    FYI we have occasionally seen similar events where, after the test, a frisky, happy implementer goes out with a press release saying "yay, we passed the test, our product rocks and did better than others".  It was probably misplaced enthusiasm, not meanspirited, but the net effect was somewhat misleading -- given that the test wasn't structured as a conformance test, and that different companies brought a broad range of different builds to the table -- some big, some small, some early, some multifunctional.
    So, I suggest you consider the following approach, based on what's worked well in the past.  These are "best practices" suggestions, not imperatives from our rules.  Some are pretty close to Diane's suggestions in her message below.  Using them would put you in a pretty good position to clear up any misunderstandings later.
    1.   Ask each participating company (note, "ask") not to issue a press release based on the test.  You might also consider asking them to agree on a joint release, or a nonbranded release from the TC chairs, at the end simply announcing the success of the informal trials.  Our Carol Geyer is experienced at getting PR people to cooperate, and could be very helpful to you here.
    2.   Adopt a statement up front saying something like, "This is a set of informal trials for development purposes, and is not a conformance test, nor a complete test of spec functionality.  It does not necessarily use finalized elements of the TC's specification, which is still in development.  Members are encouraged to employ early-stage builds in order to test both the spec and their implementations.  Our only public announcement at its conclusion will be (a) that preliminary internal tests were completed and (b) a list of participating members if they wish to be included."
    3.  Pick a subset of functionalities or cases to test that is helpful, but conspicuously does not include everything.

    Separate issue:  The question of source code, et, is potentially a little tougher, and I will write you separately on this.  I may not get that message done in time for today's meeting though.
    Regards  Jamie

~   James Bryce Clark
~   Manager Tech Stds Dev, OASIS
~   +1 978 667 5115 x 203 central office
~   +1 310 293 6739 direct

At 06:46 AM 11/10/2003, Diane Jordan wrote:
Just a reminder of our call today - WS BPEL TC implementation subgroup call  877-988-8393, int'l 203 566-8013, pc 484868.
I've taken a pass at the guidelines for a showcase at the next face to face as we discussed - here's a preliminary draft:
---- the purpose is to exercise the spec and provide validation/expose problems in it to support the TC work
---- implementations will not become part of the BPEL specification
---- it is a private event for the TC.  Correspondance and comments on it will be kept within the TC.
---- no results will be published.
---- it is intended to be a work session and not a demo.  Implementations may not be complete, work may be required to complete scenarios, etc.
---- we are planing to do some prep work beforehand (to the extent possible) so as to take best advantage of the time together.   This will at least involve providing scenarios to be attempted ahead of time and may include hosting some web based interactions.
---- it should be a friendly, collaborative event.
---- implementation may be prototypes or proofs of concept.  They do not need to be products and there is no assumption that any will become products.
Implementations will be for the use of the TC for the above purpose, and there will be no licenses to use the implementations outside the TC or for any other purpose.
---- confidential information (eg, source code and other implementation details) should not be disclosed.
---- implementations should be of the "current" spec that the TC is working on.  (The implementation sub group will figure out what exact version this will be).  Experimentation with alternate solutions to open issues may be done.
I've asked the OASIS staff for input as well.

We can discuss these and Rania's proposed scenarios on the call today.
Regards, Diane
IBM  Dynamic e-business Technologies
(919)254-7221 or 8-444-7221, Mobile: 919-624-5123

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]