Re: Different composition(or invocation) order may result an undesirable Feature Interaction problem in Web service

Hi Ray,

I am a PHD student and one of the problems I have addressed in my thesis 
is how to avoid Feature interaction in service composition. We have 
introduced the notion of compositionality, given a theoretical framework 
using Temporal Logics and also an implementation model for 
validation/verification of services at runtime to avoid undesirable 
interactiosn between services. We have also extended our model to 
account for semantic web services. Please find belwo references to 
couple of papers

Do not hesistate to get in touch for further queries.

-Monika
Monika Solanki, Antonio Cau, Hussein Zedan. Augmenting Semantic Web 
Service Description With Compositional Specification. The 13 th 
International World Wide Web Conference (Refereed Semantic Web 
track)WWW 2004  ACM, NYC- USA May 17-22, 2004
http://www.cse.dmu.ac.uk/~monika/Papers/Mypapers/313-solanki.pdf

Monika Solanki, Antonio Cau, Hussein Zedan. Introducing Compositionality 
in Web service Descriptions. 10th International Workshop on Future 
Trends in Distributed Computing Systems  FTDCS 2004 IEEE, Suzhou, 
China, May 26-28, 2004, 7 pages
http://www.cse.dmu.ac.uk/~monika/Papers/Mypapers/solankim_compositionality.pdf

Steve Ross-Talbot wrote:

>
> Interesting discussion given this is on the choreography public 
> mailing list.
>
> The Choreography approach is by design. That is you design a global 
> model that contractually prevents unwanted FI. So in a sense it is 
> impossible to get unwanted FI because all of the valid interaction are 
> modeled. This works for many real problems in which the parties 
> concerned wish to interact to do some business transactions. For 
> example most if not all financial service xaction are done this way. 
> Hence the use of vertical standards. What we do in Choreo (through 
> WS-CDL) is provide a standard way of describing such interaction from 
> a neutral perspective thus ensuring peer-to-peerness is maintained (so 
> no orchestration is needed) and providing a sound description language 
> that enables such "global-models" to be defined.
>
> Take a look at FIX (www.fixprotocol.org) as a real example of such a 
> protocol.
>
> Cheers
>
> Steve T
>
> On 29 Sep 2004, at 22:15, rayluo wrote:
>
>> Hi Chiusano,
>>
>> Thanks for your reply. Yes, sometimes the undesirable FI (Feature 
>> Interaction) can be avoided if the consumer and provider have enough 
>> "awareness". But how to achieve it? That is one of the purpose of my 
>> thesis: try to collect FI examples as many as possible,categorise and 
>> document them. Hope to offer a reference for web service designer and 
>> developer in the future. I have spent 15 months to explore those 
>> kinds of examples. But not enough examples can be identified.
>>
>> Thanks,
>>
>> Ray
>
>
>
>

-- 
**>><<**>><<**>><<**>><<**>><<**>><<**>><<**
Monika Solanki
Software Technology Research Laboratory(STRL)
De Montfort University
Gateway building, G4.61
The Gateway
Leicester LE1 9BH, UK

phone: +44 (0)116 250 6170 intern: 6170
email: monika@dmu.ac.uk
web: http://www.cse.dmu.ac.uk/~monika
**>><<**>><<**>><<**>><<**>><<**>><<**>><<**

Received on Thursday, 30 September 2004 14:34:39 UTC