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

 


Help: OASIS Mailing Lists Help | MarkMail Help

bt-spec message

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


Subject: [bt-spec] RE: [but-spec] Interoperability


Geoff

There are conformance levels in the spec based on "roles" that you can
implement  - for instance "Participant" or "Atomic Superior". I would like
to see vendors test the interoperability of theses roles across vendor
implementations. I think this would start to show where the spec is
ambiguous and could do with "tightening up". This is what I would recommend
we try and do initially. To do this we would need some test scenarios based
on the requirements and use cases we specified initially - although these
may rather out of date in light of where the spec now is.

I have liked the approach taken at the WS-I in terms of coming up with
scenario and then a profile for the scenario. For example (off the top of my
head, so don't take this to be a literal example);

Scenario: Third Party Coordination
Roles: Initiator/Terminator, Atomic Hub, Participant's)
Description: An initiator of a transaction contacts the Atomic Hub to
ascertain context for a transaction. Subsequently the initiator invokes
services and propagates the context, including the address of the Atomic
Hub. Participants enrol with the Atomic Hub. Once Context Replies have been
received from the services invoked the initiator confirms the transaction
with the Atomic Hub. The atomic Hub terminates the transaction of behalf of
the initiator, reporting back the result of the confirmation request.

Profile: Defines the interactions in terms of the specification in an
unambiguous way, in essence definitively defining, for this scenario the
interaction in term of MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT"
only, and avoids "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and
"OPTIONAL".

We may not be able to go the "whole hog" initially with defining profiles
but we should at a minimum define the scenarios and test cases to prove
conformance and therefore interoperability. I think we have dropped the ball
on this a little and got this close to completing the spec without some of
this in place, but then we are all busy and it hard to get all this done and
spec out in timely fashion. FYI WS-I has set a target of 180 days for the
basic profile that covers a simple set of scenarios  and a single profile
(covering HTTP 1.1, SOAP 1.1, WSDL 1.x, and UDDI x.x.)

Regards,
Mark Potts

PS - for Sanjay's benefit! I too hate the use of the term Hub - could we not
make it Controller!


> -----Original Message-----
> From: Geoffrey Brown [mailto:Geoffrey.Brown@oracle.com]
> Sent: Thursday, May 02, 2002 1:48 PM
> To: Mark Potts
> Cc: 'BT - spec'; William Z Pope (E-mail); Bill Flood (E-mail)
> Subject: Re: [bt-spec] Interoperability
>
>
> Mark,
>
> I agree, this makes a lot of sense .. other than issue 89
> could you elaborate on
> your thoughts ?
>
> Geoff.
>
> Mark Potts wrote:
>
> > Geoff
> >
> > With respect issues 89 is a parallel issues to ensuring
> interoperability, we
> > need to address / test interoperability between conformance
> profiles based
> > on the 1.0 spec even if it does not include serialized state.
> >
> > In terms of looking at interoperability between different
> implementations of
> > the conformance roles we need to sure up the SHOULD, MAY,
> and at least
> > define a minimum profile such that interoperability can be
> achieved between
> > conformant implementations of that profile.
> >
> > Regards,
> >
> > Mark Potts - Chief Technology Officer
> > Talking Blocks
> >
> > Office : +1 415 395 9872  x2250
> > Fax : +1 415 395 9777
> > Cell : +1 415 606 9096
> > Email :  mailto:mark.potts@talkingblocks.com
> >
> > > -----Original Message-----
> > > From: Geoffrey Brown [mailto:Geoffrey.Brown@oracle.com]
> > > Sent: Thursday, May 02, 2002 11:28 AM
> > > To: Mark Potts
> > > Cc: 'BT - spec'; William Z Pope (E-mail); Bill Flood (E-mail)
> > > Subject: Re: [bt-spec] Interoperability
> > >
> > >
> > >  Interoperability *HAS* to be addressed before v1.0 of the
> > > specification is
> > > finished, as you are aware I am trying to address this in
> > > part with issue 89.
> > >
> > > Geoff.
> > >
> > > Mark Potts wrote:
> > >
> > > > Bill,
> > > >
> > > > Originally when the TC was set up a sub-committee was formed for
> > > > Interoperability, Conformance Testing, headed up by Mark
> > > Hale at Interwoven.
> > > > I know Mark has since moved on from Interwoven and at this
> > > time is not a
> > > > contributing member to the TC.
> > > >
> > > > I suggest we address this sub-group - their charter etc in
> > > Newcastle.
> > > >
> > > > Regards,
> > > >
> > > > Mark Potts - Chief Technology Officer
> > > > Talking Blocks
> > > >
> > > > Office : +1 415 395 9872  x2250
> > > > Fax : +1 415 395 9777
> > > > Cell : +1 415 606 909
> > > > Email : mailto:mark.potts@talkingblocks.com
> > >
>



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


Powered by eList eXpress LLC