W3C home > Mailing lists > Public > public-sws-ig@w3.org > October 2004

Re: service VS profile specialization

From: David Martin <martin@AI.SRI.COM>
Date: Thu, 14 Oct 2004 23:59:25 -0700
Message-ID: <416F754D.9050808@ai.sri.com>
To: Olivier Dameron <dameron@smi.stanford.edu>
CC: public-sws-ig@w3.org



Olivier Dameron wrote:

> Hello,
>    I am not sure what is the best practice for modelling a kind of
> services (e.g. BankingService that would subsume Deposit, Withdrawal or
> Loan)
> 
> Alternative 1:
> ---------------
> BankingService is a subclass of service:Service
> No constraint is specified on the profiles presented by BankingService
> 
> Alternative 2:
> ---------------
> BankingProfile is a subclass of profile:Profile
> We only create an instance of serviceService, that presents some
> instances of BankingProfile (and possibly some other instances of
> profile:Profile)
> This is the approach taken by BravoAir
> 
> Alternative 3:
> ---------------
> Combine the previous two by defining BankingService as a subclass of
> service:Service, and BankingProfile as a subclass of profile:Profile.
> Moreover, we enforce that a banking service must present at least one
> banking profile (and respectively that a DepositService must present at
> least one DepositProfile...)
> 
> 
> The last alternative seems to be the most complete, but requires to
> maintain some similar taxonomical hierarchies for the service and the
> profile (e.g. DepositService < BankingService; WithdrawalService <
> BankingService; LoanService < BankingService; DepositProfile <
> BankingProfile; WithdrawalProfile < BankingProfile; LoanProfile <
> BankingProfile).
> So, is the service taxonomy useful, or are service just meant to be
> container for their profiles, their model and their groundings ?

There are no hard-and-fast rules; that is, this is largely a matter of 
style.

But in my view there is no need for a service taxonomy.  As you say, 
services are just meant to be containers for their profiles, their model 
and their groundings ?  I recommend Alternative 2.

Regards,
David
Received on Friday, 15 October 2004 06:59:45 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Sunday, 16 March 2008 00:10:58 GMT