Sure you are changing legacy code: you are adding the "transducer proxy". How do you know it does not disrupt existing deployments? This is not something up to you or me to decide.
Ugo
> -----Original Message-----
> From: Satish Thatte [mailto:]
> Sent: Friday, November 21, 2003 12:34 PM
> To: Ugo Corda; Sanjiva Weerawarana; Francisco Curbera
> Cc: Ron Ten-Hove;
> Subject: RE: [wsbpel] Issue - 77 - Under specified operation
> definitions
>
>
> No need to change legacy code. It is perfectly possible to
> build a tool
> to take one of the "legacy WSDL" definitions, inspect the binding
> element and generate a "better designed abstract WSDL" from it
> automatically, along with an appropriate "transducer proxy". Might be
> good business too if the problem is common enough, if not why would we
> even bother to think about it? :-)
>
> -----Original Message-----
> From: Ugo Corda [mailto:]
> Sent: Friday, November 21, 2003 12:24 PM
> To: Sanjiva Weerawarana; Satish Thatte; Francisco Curbera
> Cc: Ron Ten-Hove;
> Subject: RE: [wsbpel] Issue - 77 - Under specified operation
> definitions
>
> Ok, I think we are making progress understanding what exactly
> WSIF does.
>
> > The client of the service
> > (who's the one using WSIF) only gets to use the service thru the
> > service's interface (portType). That's the guiding spirit and
> > principle of WSIF.
>
> Fine, so the WSIF client would not be able to manipulate the "Header"
> abstract message in my example (in other words, WSIF behaves the same
> way BPEL does).
>
> As I pointed out before, other WS frameworks do not behave
> that way, and
> they allow users to directly manipulate the "Header" abstract
> message in
> my example (I know for sure because a customer of ours brought us that
> type of example asking us to support it in BPEL).
>
> You might say that using WSDL 1.1 that way is a bad idea. I am not
> arguing with that. All I am saying is that it is perfectly
> legitimate to
> use it that way according to WSDL 1.1 (the spec actually explicitly
> calls out that case in sec. 3.7, when it says "The referenced message
> need not be the same as the message that defines the SOAP body").
>
> So some users decided to use WSDL 1.1 that way, and we cannot
> tell them
> they made a mistake (they didn't: they just followed the spec). So we
> still have to deal with the legacy issue presented by these
> implementations.
>
> How you solve this legacy issue is evidently very much related to your
> company's business strategies and priorities. Some companies
> tell their
> customers they should rewrite their code when it is not "right" (and
> they are happy to help them do that - for a fee, of course). My own
> company happens to work in a different type of business,
> which is based
> on making legacy applications interoperate. So not asking our
> customers
> to modify their existing code is one of our business' highest
> priorities
> ...
>
> Ugo
>
> > -----Original Message-----
> > From: Sanjiva Weerawarana [mailto:]
> > Sent: Friday, November 21, 2003 12:01 PM
> > To: Ugo Corda; Satish Thatte; Francisco Curbera
> > Cc: Ron Ten-Hove;
> > Subject: Re: [wsbpel] Issue - 77 - Under specified operation
> > definitions
> >
> >
> > OK let me try .. I'm not sure what you mean by a WSIF application,
> > but an app that's using a WSDL to invoke a service should only
> > be looking at the portType of the service. If the binding refers
> > to other stuff (other abstract messages, whatever), then that's
> > the business of the binding implementation. The client of
> the service
> > (who's the one using WSIF) only gets to use the service thru the
> > service's interface (portType). That's the guiding spirit and
> > principle of WSIF.
> >
> > I wouldn't characterize that as "WSIF only supports part of WSDL."
> > I'd say that WSDL v1.1 had a grey area in the SOAP binding element
> > soap:header and other WSIF bindings may have the same - basically
> > that where that optional message attribute (of soap:header, for
> > example) gets an actual message from is undefined. The general
> > feeling (at least on the IBM side) was that that comes from your
> > context, but of course its not documented. As such using it is
> > simply a bad idea and in any case WSDL 2.0 will likely not have
> > such a mechanism.
> >
> > Sanjiva.
> >
> > ----- Original Message -----
> > From: "Ugo Corda" <>
> > To: "Satish Thatte" <>; "Francisco Curbera"
> > <>
> > Cc: "Ron Ten-Hove" <>; "Sanjiva Weerawarana"
> > <>; <>
> > Sent: Friday, November 21, 2003 11:04 PM
> > Subject: RE: [wsbpel] Issue - 77 - Under specified operation
> > definitions
> >
> >
> > > No, actually it wasn't. It was not clear whether Paco
> > referred to the
> > soap:header binding or the abstract "Header" message. Since
> > he said "or
> > anything else in the binding" I would assume he was only
> > referring to the
> > binding part.
> > >
> > > In any case, I don't want to waste to much time
> discussing what WSIF
> > exactly does (this was just an example, which might or might not be
> > relevant). There are only two possibilities here:
> > >
> > > 1. A WSIF application cannot manipulate the "Header"
> > abstract message.
> > > Ok, so WSIF cannot handle the full range of WSDL files
> > allowed by WSDL
> > 1.1. Sorry, my WSIF example was not relevant to this
> > discussion. (Of course,
> > that does not mean other frameworks cannot handle it either -
> > in fact, my
> > example in issue 77 is taken from a real existing application
> > of a customer
> > of ours).
> > >
> > > 2. A WSIF application can manipulate the "Header"
> abstract message.
> > > In this case the WSIF example is relevant and shows that
> > WSIF applications
> > can handle what BPEL cannot.
> > >
> > > Ugo
> > >
> > > > -----Original Message-----
> > > > From: Satish Thatte [mailto:]
> > > > Sent: Friday, November 21, 2003 8:52 AM
> > > > To: Ugo Corda; Francisco Curbera
> > > > Cc: Ron Ten-Hove; Sanjiva Weerawarana;
>
> > > > Subject: RE: [wsbpel] Issue - 77 - Under specified operation
> > > > definitions
> > > >
> > > >
> > > > Hmm.. Paco wrote
> > > >
> > > > The application using WSIF to access the service does not see
> > > > the header
> > > > message or anything else in the binding
> > > >
> > > > Was that not clear enough?
> > > >
> > > > -----Original Message-----
> > > > From: Ugo Corda [mailto:]
> > > > Sent: Friday, November 21, 2003 8:45 AM
> > > > To: Francisco Curbera
> > > > Cc: Ron Ten-Hove; Sanjiva Weerawarana;
>
> > > > Subject: RE: [wsbpel] Issue - 77 - Under specified operation
> > > > definitions
> > > >
> > > > Paco,
> > > >
> > > > The question relevant to this discussion is the
> following: can the
> > > > application using WSIF see (i.e. set or get the value of)
> > the abstract
> > > > message defined outside the abstract port? We know that
> > BPEL currently
> > > > cannot.
> > > >
> > > > Ugo
> > > >
> > > > > -----Original Message-----
> > > > > From: Francisco Curbera [mailto:]
> > > > > Sent: Friday, November 21, 2003 8:22 AM
> > > > > To: Ugo Corda
> > > > > Cc: Ron Ten-Hove; Sanjiva Weerawarana;
> >
> > > > > Subject: RE: [wsbpel] Issue - 77 - Under specified operation
> > > > > definitions
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > > Hi Ugo,
> > > > >
> > > > > I would not agree with that conclusion. The way WSIF would
> > > > > support your
> > > > > example in issue 77 is through a binding module (called a
> > > > > "provider" in
> > > > > WSIF) that understans and takes care of the header as
> > > > > declared by the WSDL
> > > > > binding. The application using WSIF to access the service
> > > > > does not see the
> > > > > header message or anything else in the binding. Same thing
> > > > > with non-SOAP
> > > > > bindings. This is a key design point because otherwise you
> > > > loose your
> > > > > ability to use different access channels with different QoS
> > > > > characteristics
> > > > > to reach the same service (different bindings). It also
> > results in a
> > > > > cleaner programming model.
> > > > >
> > > > > Regards,
> > > > >
> > > > > Paco
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > > "Ugo Corda"
> > > > >
> > > > >
> > > > > <UCorda@SeeBeyond To: "Ron
> > > > > Ten-Hove" <>, "Sanjiva Weerawarana"
> > > > >
> > > > > .com>
> > > > > <>
> > > > >
> > > > > cc:
> > > > > <>
> > > > >
> > > > > 11/20/2003 01:43 Subject: RE:
> > > > > [wsbpel] Issue - 77 - Under specified operation definitions
> > > > >
> > > > > PM
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > > I conclude from this that my WSIF example seems
> > appropriate for this
> > > > > discussion. WSIF would be able to support a case like the one
> > > > > I gave for
> > > > > issue 77, and it would be able to map the "Header" abstract
> > > > > message to any
> > > > > particular binding I want to express (including, but not
> > > > > limited to, SOAP -
> > > > > in particular, bindings that don't even have the concept of a
> > > > > header)). Is
> > > > > that correct?
> > > > >
> > > > > Ugo
> > > > > -----Original Message-----
> > > > > From: Ron Ten-Hove [mailto:]
> > > > > Sent: Thursday, November 20, 2003 10:30 AM
> > > > > To: Sanjiva Weerawarana
> > > > > Cc:
> > > > > Subject: Re: [wsbpel] Issue - 77 - Under specified operation
> > > > > definitions
> > > > >
> > > > > WSIF is a set of bindings for Java and J2EE; it doesn't
> > > > > extend WSDL in any
> > > > > non-standard way that I am aware of. WSIF "understands" WSDL
> > > > > files that use
> > > > > those bindings. The message model is plain WSDL 1.1.
> > > > >
> > > > > -Ron
> > > > >
> > > > > Sanjiva Weerawarana wrote:
> > > > > I'm confused .. can you give an example of what
> > you mean by
> > > > > "abstract messages that are not part of an abstract
> > > > operation"?
> > > > > Maybe I haven't understood what you have in mind ..
> > > > >
> > > > > Thanks,
> > > > >
> > > > > Sanjiva.
> > > > >
> > > > > ----- Original Message -----
> > > > > From: "Ugo Corda" <>
> > > > > To: "Sanjiva Weerawarana"
> > <>; "Francisco
> > > > > Curbera"
> > > > > <>
> > > > > Cc: "Ron Ten-Hove" <>;
> > "Satish Thatte"
> > > > > <>; <>;
> > > > > <>
> > > > > Sent: Thursday, November 20, 2003 7:04 AM
> > > > > Subject: RE: [wsbpel] Issue - 77 - Under
> > specified operation
> > > > > definitions
> > > > >
> > > > >
> > > > >
> > > > > So are you saying that WSIF can only process a
> > > > > subset of all
> > > > > the legal
> > > > >
> > > > > (according to WSDL 1.1) WSDL files? (Please notice
> > > > that I am not
> > > > > talking
> > > > > about adding any extension at the abstract level - just
> > > > > supporting
> > > > > what is
> > > > > allowed by WSDL 1.1)
> > > > >
> > > > > Ugo
> > > > >
> > > > >
> > > > > -----Original Message-----
> > > > > From: Sanjiva Weerawarana
> > > > > [mailto:]
> > > > > Sent: Wednesday, November 19, 2003 5:00 PM
> > > > > To: Ugo Corda; Francisco Curbera
> > > > > Cc: Ron Ten-Hove; Satish Thatte;
> > > > > ;
> > > > >
> > > > > Subject: Re: [wsbpel] Issue - 77 -
> > Under specified
> > > > > operation
> > > > > definitions
> > > > >
> > > > >
> > > > > "Ugo Corda" <> writes:
> > > > >
> > > > > Just curious: does WSIF allow
> > you to define
> > > > > abstract
> > > > >
> > > > > messages that are not
> > > > > part of an abstract operation?
> > > > >
> > > > > Ugo
> > > > >
> > > > > No it doesn't; WSIF only adds additional
> > > > bindings to
> > > > > WSDL,
> > > > > not anything
> > > > > at the abstract level.
> > > > >
> > > > > Sanjiva.
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > > To unsubscribe from this mailing list (and be removed
> > > > > from the roster
> > > > > of the OASIS TC), go to
> > > > >
> > > > > http://www.oasis-open.org/apps/org/workgroup/wsbpel/members/le
> > > > ave_workgroup.php
> > > > .
> > > >
> > > >
> > > >
> > > >
> > > > To unsubscribe from this mailing list (and be removed from
> > > > the roster of
> > > > the OASIS TC), go to
> > > > http://www.oasis-open.org/apps/org/workgroup/wsbpel/members/le
> > > ave_workgr
> > > oup.php.
> > >
> > >
> > >
> > >
> > > To unsubscribe from this mailing list (and be removed from
> > the roster of
> > the OASIS TC), go to
> > http://www.oasis-open.org/apps/org/workgroup/wsbpel/members/le
> ave_workgroup.
> php.
>
>
>
>
