RE: Web Service Description and stateful services - (on the 'www -ws@w3.org' list) Debating on a) Stateful Web Service Instances b) Sta teful Interaction - OGSI

As an ex-physicist, I cannot help but summarising the following "service
law" after observing people's "virtual experiments" on this list.

The Law of Stateful service:

"How stateful a service is is determined by the amount of customisation the
service delivers."

When there is zero customisation, the service is totally stateless. As soon
as the service starts differentiating its requestors, it becomes stateful.
Interestingly, a service can become more stateful as it has more
interactions with a requestor. That is why you like to go to your favourite
restaurant. 

So what have we learned here?  Stateful services are useful and important
because they provide added values to their customers.  How to provide
stateful service is really service provider (application) specific.  What
can we standardise then?

In order to provide stateful service, the service provider must be able to
distinguish its service requestors. So, we just need to define a standard
way of providing a requestor differentiator. 

Hao

-----Original Message-----
From: Ugo Corda [mailto:UCorda@SeeBeyond.com]
Sent: Sunday, June 22, 2003 2:56 AM
To: Newcomer, Eric
Cc: www-ws-arch@w3.org
Subject: RE: Web Service Description and stateful services - (on the
'www-ws@w3.org' list) Debating on a) Stateful Web Service Instances b)
Stateful Interaction - OGSI



> I think semantics and state are different things.  You can have either
without the other.

Semantics is always a vague concept, so let me try to further clarify. What
I was referring to is something that describes the internal behavior of the
service and is currently not expressed in the WSDL interface. To take Chris'
example, if a client sends a message containing two coordinates to a map
service with the expectation that the service will refer to a map that the
client previously retrieved, this particular server behavior is what I would
say to be part of its semantics (i.e. not expressed in the interface).

Ugo

Received on Tuesday, 24 June 2003 01:27:49 UTC