Re: question: relation between WS-CDL and WSCI

Fair question Titi.

Okay where do I start....
First of all WSCI was an executable language. Secondly WSCI has no 
formal basis at all.
When we started Web Services Choreography we started from requirements 
and spent a long time thinking about what was needed as opposed to what 
was there. We came up with a list of requirements and a list of 
objectives. All these are listed in the requirements document on the 
Choreography WG website. WSCI and WSCL and others did not meet these 
requirements at all. So to be able to meet the requirements we really 
had no option but to start from scratch.

One thing did cross the boundary into, namely the concept of the 
"global model". This remains a principle focus of the approach we have 
taken. We tightened up by basing things around some formalism (the 
pi-calculus). We recruited some academic experts in the field (Robin 
Milner, Nobuku Yoshida and Kohei Honda) to help make this real. And we 
issued WS-CDL (editors draft and overview document).

We did embark on a formal look at shortcomings, which you can see from 
previous minutes, but this work never amounted to anything.

I hope this helps put some context around things. Of course when the 
group was set up a lot of politics was flying around regarding WSCI and 
BPEL4WS. What we did was expose an area that no one was really looking 
at and have since found a lot of commercial traction with the global 
model and the idea of a global model led contract based design.

Interestingly enough if you look at financial services trading systems, 
they all work based on an informal global model which described the 
external observable behaviour of a multi-party protocol. This is 
exactly what WS-CDL does but in a formal setting that can make 
guarantees about certain liveness properties and conformance. It turns 
out this is exactly what these folks need. I suspect that this is not 
the only area that needs such a mechanism. It would appear that this is 
how these vertical standards really do work. They all use some form of 
informal global model or multiparty contract couples with data formats. 
Methinks this applies outside this area into telco, SCM and so on.

Cheers

Steve T


On 12 Apr 2004, at 14:44, Titi Roman wrote:

>
> Hi,
> Thanks for your answer.
> I still have a question. When you say that it would be better for the
> community at large to forget about WSCL and WSCI, I suppose that they 
> had
> many technical shortcomings. Could you explain in a few words these
> shortcomings or point some documents where I can find these 
> shortcomings
> briefly explained?
> Why W3C Choreography group chose to build another choreography language
> instead of extending WSCI (or other initiatives you pointed in your 
> email)?
> Could you also give me a pointer to the charter you were talking about 
> in
> you email (where the relation between WS-CDL and WSCI is described)?
> Thanks a lot,
> Titi Roman
>
>
> ----- Original Message -----
> From: Steve Ross-Talbot <steve@enigmatec.net>
> To: Titi Roman <dumitru.roman@deri.ie>
> Cc: WS-Choreography List <public-ws-chor@w3.org>
> Sent: Sunday, April 11, 2004 3:39 PM
> Subject: Re: question: relation between WS-CDL and WSCI
>
>
>>
>> Titi,
>>
>> The relationship is described in our charter (). WSCI will not 
>> progress
>> further. WS-CDL will progress further. WSCI was the basis of a W3C
>> working note that preceeded the formation of the W3C Web Services
>> Choreography Working Group. It forms an input into the work of the
>> working group but does not define it nor restrict it. WSCL was also a
>> working note and BPML, DAML/S and BPEL4WS were all listed as inputs.
>> The only difference is that WSCI and WSCL are W3C working notes and 
>> the
>> others are not (maybe not true in the case of DAML/S).
>>
>> I hope that explains it. It would be better for the community at large
>> to forget about WSCL and WSCI. Unless you are doing a history
>> assignment they have very little of any relevance anymore.
>>
>> Cheers
>>
>> Steve T
>>
>> On 10 Apr 2004, at 14:51, Titi Roman wrote:
>>
>>> Hi,
>>> I would like to ask you which is the relation between the WS-CDL and
>>> WSCI 1.0 (I even see a reference ([15]) in WS-CDL editor's
>>> draft(http://www.w3.org/2002/ws/chor/edcopies/cdl/cdl.html) which is
>>> never used)?
>>> What was wrong with the WSCI such that a new choreography language
>>> (WS-CDL) was needed?
>>> Thank you,
>>> Titi Roman
>>
>>

Received on Sunday, 11 April 2004 16:48:45 UTC