W3C home > Mailing lists > Public > public-ws-chor@w3.org > April 2004

RE: question: relation between WS-CDL and WSCI

From: Jean-Jacques Dubray <jeanjadu@Attachmate.com>
Date: Sat, 17 Apr 2004 07:21:19 -0700
Message-ID: <D15269CBED76D51185280008C73323FA02CE0DB6@exch-bel6.attachmate.com>
To: "Titi Roman" <dumitru.roman@deri.ie>, "Nickolas Kavantzas" <nickolas.kavantzas@oracle.com>
Cc: "WS-Choreography List" <public-ws-chor@w3.org>

Titi:

I think that you are absolutely correct in saying that a "WS-behavior
description language is needed". As pointed out by WSCI, WSDL only
provides a description of the ports of a service but not the way (e.g.
sequence) once should use these ports if they have a correlated usage
pattern. 

I also think that Ugo is correct in saying that abstract BPEL fulfills
this role, as did WSCI. This information should be used to validate that
a given service can participate in a given choreography. So the users of
that service which define the choreography can validate that this
particular service will operate accordingly.

That being said, this WSDL# (WSCI, abstract BPEL) is not a choreography
definition, nor, should a choreography definition reflect it. As you
mention this service can work with any number of choreographies that is
conforms too. What you need is really an INTERFACE description language.
From a choreography, the only thing you can do is generate a WSDL#
skeleton but probably hundreds of abstract BPEL could fit a given
choreography. You can also generate ONE choreography from a WSDL#, but
again, thousands of choreographies could fit that particular WSDL#.

I think that your question is precisely pointing the need for all
concepts:
- choreography (WS-CDL)
- behavioral interface description (abstract BEPL)
- implementation of a given interface (BPEL)

Note that abstract BPEL is not always mandatory if you have a BPEL
service and want to see if it conforms to a choreography. However, when
the choreography is unknown at design time, and you want this service to
be used by many different consumers (an not expose the intricacies of
your service), then an abstract BPEL (or WSCI) would be the right way to
go.

Feel free to cross post that on the BPEL list if you see fit.

JJ-

-----Original Message-----
From: Titi Roman [mailto:dumitru.roman@deri.ie] 
Sent: Friday, April 16, 2004 8:02 PM
To: Nickolas Kavantzas
Cc: WS-Choreography List
Subject: Re: question: relation between WS-CDL and WSCI


Hi Nick,
Thanks for your answer. See just some questions/comments in-line that
just
crossed my mind.

----- Original Message -----
From: Nickolas Kavantzas <nickolas.kavantzas@oracle.com>
To: Titi Roman <dumitru.roman@deri.ie>
Cc: Assaf Arkin <arkin@intalio.com>; WS-Choreography List
<public-ws-chor@w3.org>
Sent: Thursday, April 15, 2004 8:14 PM
Subject: Re: question: relation between WS-CDL and WSCI

>
> See comments about WS-CDL below.
>
> --
> Nick
>
> >
> > As far as I understood and please correct me if I am wrong, WS-CDL
supports
> > only "global model"(i.e. the multi-participant view of the overall
message
> > exchange) - this sounds to me as a static linking between web
services.
What
> > about the dynamic linking between ws from the point of view if
> > coneversation?
>
> Why does the global model implies static linking? I am not sure I
understand this.
>
> And what is the dynamic linking between ws from the point of view of
the
> coneversation? Can you elaborate on this more?

What I meant by this was that when you specify a choreography all the
participants are known a-priori (static linking).
What I miss is: I want to make a service accessible through the internet
and
I want to describe in a language (hopefully WS-CDL) how it behaves when
someone uses it - note that I do not need a "global model" to do this
since
I do not care who will use it (e.g. a "credit check" service will offer
its
service to potentially thousands of travel agents - it is not the job of
the
"credit check" service to conform to the thousands choreographies of the
travel agents, but is the job of the travel agents to discover the
"credit
check" service and adapt their choreographies to the "non-global model"
choreography of the "credit check" service (i.e. dynamic linking)) - is
it
possible to describe the way "credit check" service behaves in my
scenario
in WS-CDL?
if yes, I personally find it very unintuitive in this, I suppose, very
realistic scenario(I understood from your mail that you and Steve have a
proposal related to this issue - is it possible for me to have access to
it?)
As you can see my example is in the context of section "3.1.1.3.2
Variation
2" of your current draft on requirements - while it can be the case for
the
"credit check"  service (new credit bureau) to "join" the travel agent
choreography, it can also be the case (and I personally believe that it
is
very realistic) for the travel agent to join the "non-global model"
choreography of the "credit check" service(which I tried to point out in
the
above example).
Then why not just add another requirement(e.g. "A sub-choreography must
be
generated and integrated in the "global model" choreography for each
service
discovered and used o the fly") and try to tackle also this, I believe
very
natural problem, in the choreography language?

Thanks a lot and I will surely read carefully the other comments from
you
and Steve's last email and I'll come up with new questions ;-).
Titi Roman

>
> > What about the view of the overall message exchange as seen
> > from one participant? Isn't WS-CDL supposed to support also this
view?
> >
>
> I have filled an issue for this in the Cambridge F2F and myself and
Steve
have
> an action item to present our proposal to the group.
>
Received on Saturday, 17 April 2004 10:24:00 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 19:30:24 UTC