RE: [RTF] Behavior definition of Services - public discussion

Suresh,

>Another solution: every service defn may also carry a defn of what is the
>equivalence behavior.

I have a hard time imagining how this defn would look like. Would it be part
of a WSDL definition? Would it be a URL pointing to a human-processable
document? Would it be something that could be machine-verifiable?

For me the bottom line is that semantics is a quicksands area for computing
in general (having had direct experience of the AI disillusion of the
eighties) and for Web Services in particular. I would be very cautious
before introducing concepts like semantic equivalence that could create
expectations well beyond what Web Services can actually deliver today.

From a political point of view, it seems to me that the safest way to
proceed in this area is by pointing to work being done in other groups ;).

Ugo

-----Original Message-----
From: Damodaran, Suresh [mailto:Suresh_Damodaran@stercomm.com]
Sent: Tuesday, July 02, 2002 4:56 PM
To: 'Ugo Corda'; 'www-ws-arch@w3.org'
Subject: RE: [RTF] Behavior definition of Services - public discussion


Ugo,

My comments are inline.

-Suresh
Sterling Commerce   



-----Original Message-----
From: Ugo Corda [mailto:UCorda@SeeBeyond.com]
Sent: Tuesday, July 02, 2002 6:34 PM
To: Damodaran, Suresh; Ugo Corda; 'www-ws-arch@w3.org'
Subject: RE: [RTF] Behavior definition of Services - public discussion


Suresh,

>(1) is the notion of semantic behavior equivalence defined in AC009 (I
don't believe so)

Well, I don't think we can talk about semantic behavior equivalence without
knowing how to define semantic behavior first. Unless you are thinking of
semantic behavior in terms of inputs and outputs (so that if two services
have the same inputs and outputs values they are also equivalent). But I
think semantic behavior should go much deeper than that, by taking into
account, for instance, all the side effects not accounted for by IO
parameters, and at that point it becomes much more difficult to say when two
semantic behaviors are "equivalent".

<sd> Ugo, this is the exact difficulty I faced.
Where do you draw the line on what is "equivalent." One way is to wait for
a "universal equivalence defn." to appear via Semantic Web/.... 
Another solution: every service defn may also carry a defn of what is the
equivalence behavior.
This way, we don't have to depend on Semantic web to provide a universal
"equivalence definition."

I clearly prefer the later solution. What do you think?
</sd>


Ugo

-----Original Message-----
From: Damodaran, Suresh [mailto:Suresh_Damodaran@stercomm.com]
Sent: Tuesday, July 02, 2002 4:21 PM
To: 'Ugo Corda'; 'www-ws-arch@w3.org'
Subject: RE: [RTF] Behavior definition of Services - public discussion


Ugo,

If I understand you correctly, you are proposing that we delay defining
19.2.2 below
till AC009 is well defined? I am not sure we should.

I believe the idea of "semantic behavior equivalence" is independent of how
it is achieved
- through semantic web/design by contract/enumeration (with an old fashioned
valid input-output table)/.... Therefore, importing "semantic behavior
equivalence definition" from elsewhere (may be multiple places) should be
just fine.

The question is, (1) is the notion of semantic behavior equivalence defined
in AC009
(I don't believe so)
(2) should we use it to define stable definitions 
(I think we should).

Cheers,

-Suresh
Sterling Commerce   



-----Original Message-----
From: Ugo Corda [mailto:UCorda@SeeBeyond.com]
Sent: Tuesday, July 02, 2002 5:39 PM
To: Damodaran, Suresh; 'www-ws-arch@w3.org'
Subject: RE: [RTF] Behavior definition of Services - public discussion


Suresh,
If I correctly understand the spirit of AC009, WSA's plan is to adopt
semantic technologies developed elsewhere (e.g. the Semantic Web Activity).
I would imagine that definitions of semantic behavior equivalence would be
part of such adopted technologies, so that WSA could import the specific
equivalence concepts and technologies from those external efforts. In other
words, I think it is premature to address semantic identity within WSA
before concrete links are established with external semantic activities in
the spirit of AC009.

Regards,
Ugo

-----Original Message-----
From: Damodaran, Suresh [mailto:Suresh_Damodaran@stercomm.com]
Sent: Tuesday, July 02, 2002 3:27 PM
To: 'Ugo Corda'; 'www-ws-arch@w3.org'
Subject: RE: [RTF] Behavior definition of Services - public discussion


Ugo,

May be so. If it is defined in AC009, I am interested in knowing how AC009
helps in explicit definition of service behavior + the requirement below.

Cheers,

-Suresh
Sterling Commerce   



-----Original Message-----
From: Ugo Corda [mailto:UCorda@SeeBeyond.com]
Sent: Tuesday, July 02, 2002 5:16 PM
To: Damodaran, Suresh; 'www-ws-arch@w3.org'
Subject: RE: [RTF] Behavior definition of Services - public discussion


Suresh,
Isn't semantics already addressed by AC009?

Ugo
SeeBeyond

-----Original Message-----
From: Damodaran, Suresh [mailto:Suresh_Damodaran@stercomm.com]
Sent: Tuesday, July 02, 2002 3:00 PM
To: 'www-ws-arch@w3.org'
Subject: [RTF] Behavior definition of Services - public discussion




Hi all,

While discussing reliability of web services in RTF, we hit upon the issue
of how to define the "behavior"
of a web service. Service Defn. based on WSDL only allows the interface
description, and is silent
about what the service will do (semantics). For example, a service
description that takes two parameters and does an "add"
may do a multiplication. The question is whether WSA should "enable" such
semantic definition of the behavior
of services. There may be multiple means to accomplish this, including
"design by contract"[2]. Many may argue that
such a definition may not be complete in most circumstances. In any case,
what do you think?


Here is the item tabled for debate from [1] 
D-AR019.2.2 The functional behavior of two or more web service implementing
the same version (see AR019.3.1) of a web service definition is identical.
	[<sd> the reverse may be true also - two implementations may have
the same behavior but different definitions, but is not worth mentioning
</sd>]

Thanks,

-Suresh
Sterling Commerce   
[1] http://lists.w3.org/Archives/Public/www-ws-arch/2002Jun/0186.html
[2]  http://www.eiffel.com/doc/manuals/technology/contract/

Received on Tuesday, 2 July 2002 20:17:28 UTC