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

RE: Re: Use Case Proposal - Supporting the classifier-instance relationship in Web Services

From: <marco.adragna@virgilio.it>
Date: Sat, 8 May 2004 09:59:58 +0200
Message-ID: <408712C9000200FC@ims6b.cp.tin.it>
To: "Ugo Corda" <UCorda@SeeBeyond.com>
Cc: public-ws-chor@w3.org

Thank you for your reply.
I didn't know that the ws-resource approach had been submitted to OASIS,
but I have been following the evolution of Grid Services. In early 2003
was working on my dissertation on the relationship between OO design and
Web Services. On June 2003 I was exchanging views in the WS Architecture

group mailing list on what "Stateful Web Service Instance" and OGSI should

mean to the web service community. 
Later I was following the work of the Attribute task force. In the last
months I was unable to continue following those events due to 
an increased work pressure.

I still think that WS-CDL should be the right place to address such Use
because supporting it relates to the general concept of Abstraction,
of "X is instance of the classifier Y". This is a fundamental Requirement

Engineering concept, that doesn't belong only to OO.

Have a truly nice day,

>-- Messaggio originale --
>Subject: RE: Re: Use Case Proposal - Supporting the classifier-instance
>in Web Services
>Date: Fri, 7 May 2004 15:51:58 -0700
>From: "Ugo Corda" <UCorda@SeeBeyond.com>
>To: <marco.adragna@virgilio.it>
>Cc: <public-ws-chor@w3.org>
>The Use Case you describe sounds like something that could also be
>addressed by the Web Services Resource Framework (recently submitted to
>OASIS - see
>http://www.oasis-open.org/committees/tc_home.php?wg_abbrev=wsrf ). Do
>you happen to know the WS-RF specs? If you do, do you still think there
>are reasons for preferring the WS-CDL approach?
>Thank you,
>> -----Original Message-----
>> From: public-ws-chor-request@w3.org 
>> [mailto:public-ws-chor-request@w3.org] On Behalf Of 
>> marco.adragna@virgilio.it
>> Sent: Friday, May 07, 2004 6:36 AM
>> To: steve@enigmatec.net
>> Cc: public-ws-chor@w3.org
>> Subject: Re: Re: Use Case Proposal - Supporting the 
>> classifier-instance relationship in Web Services
>> Steve,
>> Thank you for letting me know abouth the new requirement 
>> document. Sorry for responding with such delay, I wish this 
>> was part of my job but it isn't, so a brief morning session 
>> was all I could dedicate to it. On the other hand, it was a 
>> job-related need that first got me thinking to a 
>> Classifier-Instance choreography. Here I shortly describe 
>> such common business situation. I have read the new 
>> requirement document: my Use Case might contain some 
>> additional requirements. Unfortunately I am unable to 
>> formally express such reqs. All I have is a gut feeling that 
>> something might still be missing, but I am unable to point my 
>> finger to it. Thank you for taking the time of examining my 
>> not-so-clear Use Cases(!)
>> Marco Adragna
>> Senior Consultant
>> ItaliaLavoro, the technical agency of the Italian Ministry of Work.
>> A Company wants to make it easier to expose the functionality 
>> of existing OO components as XML web service. The goal is to 
>> minimise the need of writing component specific code, whilst 
>> still exposing fully functional and truly standard web 
>> services. Those components have been developed using 
>> different technologies, purchased from different vendors. 
>> Some OO components offer a very coarse grained functionality 
>> and are relevant to business processes carried out both 
>> internally and externally to the Company. Vendors provide 
>> tools capable of creating a WSDL remote interface from OO 
>> components written following certain rules. (e.g. a properly written
>> stateless session bean in J2EE or 
>> System.Web.Services.WebService class in .Net)
>> The problem is that a generic (vendor-neutral) web service 
>> client is unable to get complete information on how to use 
>> those components merely by reading such WSDL 
>> document.Furthermore, too many components can't be exposed as 
>> web services or need a lot of custom coding to do so. 
>> Components are not web services: they typically need to be 
>> instantiated before they can be used, instances need to be 
>> addressable via a form of remote reference, instances have 
>> lifetime. It is impossible to express such needs using WSDL, 
>> hence generic web service clients can't know how to correctly 
>> use those OO-component-powered Web Services.
>> The solution is writing vendor-specific tools that can 
>> "process" a generic OO component, creating 2 WSDL and a 
>> WS-CDL. Those act as a "more intelligent" remote interface. 
>> An instance of a WS-CDL described choreography provides the 
>> context in which a transient ComponentInstance can begin, do 
>> stuff and end.
>> Typically:
>> - a ComponentParent WSDL for creating, finding and lifetime-related
>>   operations.
>> - a ComponentInstance WSDL with the business logic.
>> - a WS-CDL specifying in which sequence the two should be used by one
>>   or more Users.
>> The basic steps of the interaction are listed below:
>> 1- 	User requests ComponentParent a channel to a new 
>> ComponentInstance.
>> 	(This is marked as Choreography Initiator in the WS-CDL)
>> 2- 	User interacts with ComponentInstance via the channel.
>> 3- 	Other Users might join in, obtaining the channel of the existing
>> 	ComponentInstance via lookup services provided by 
>> ComponentParent,
>> 	and then interacting with such ComponentInstance.
>> 4-	Users stop interacting with the ComponentInstance. Such specific
>> 	ComponentInstance is no longer useful. (The complete 
>> condition of
>> 	the WS-CDL evaluates to true. The Choreography ends.)
>> Marco Adragna
>> marco.adragna@virgilio.it 
>> ________________________________________________________
>> (Steve's email...)
>> >Marco,
>> >We have a new requirements document on the W3C website and wanted to

>> >check if this covered you requirements in the use case you submitted

>> >sometime ago. Could you take a look and let us know.
>> >
>> >http://www.w3.org/TR/2004/WD-ws-chor-reqs-20040311/
>> >
>> >Best Regards
>> >
>> >Steve Ross-Talbot
>> >Chair W3C Web Services Coordination Group
>> >co-Chair W3C Web Services Choreography Working Group
>> >
>> >O: +44 207 397 8207
>> >C: +44 7855 268 848
>> >www.enigmatec.net
>> >
>> >
>> >On 16 Sep 2003, at 09:23, marcoadr@tin.it wrote:
>> >
>> >>
>> >> I have a Use Case to propose for consideration.
>> >>
>> >> Composition languages allow the creation a new web service from 
>> >> existent ones.
>> >> This define an aggregation/part relationship between the new
>> >> composition
>> >> and the individual web services that compose it.
>> >> In requirement engineering both the aggregation/part and the
>> >> classifier/instance
>> >> relationships are fundamental "tools"
>> >> in structuring Problem Analysis and Behavioural requirements.
>> >> The process of turning a problem analysis into a working software
>> >> solution
>> >> can be simple only if the
>> >> software technology used to implement such solution supports the
>> >> fundamental
>> >> "tools" of problem analysis.
>> >> Web services arguably lack of a complete support for expressing
>> >> classifier/instance
>> >> relationship among them.
>> >> With the following use case, I hope to contributes to the 
>> support for
>> >
>> >> the
>> >> classifier/instance relationship in Web Services.
>> >>
>> >> Regards
>> >> Marco Adragna
>> >> ......
>> >
>> >
Received on Saturday, 8 May 2004 04:02:19 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:01:04 UTC