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

Re: [ann] CODE (CMU OWL-S Development Environment) - Beta release available

From: Bijan Parsia <bparsia@isr.umd.edu>
Date: Fri, 3 Dec 2004 00:34:50 +0900
Message-Id: <B4CE222B-4477-11D9-8B14-000D93C1F7A6@isr.umd.edu>
Cc: public-sws-ig@w3.org, Naveen Srinivasan <naveen@cs.cmu.edu>
To: Paul Libbrecht <paul@activemath.org>

On Dec 3, 2004, at 12:24 AM, Paul Libbrecht wrote:

>
> Hi,
>
> While reading this, I came to think that all these translations Java > 
> WSDL > OWL-S, though well integrated, may be somewhat "lossy" in terms 
> of expressivity.

Well, it depends on what you mean. In one sense, yes, in another sense, 
no :)

There is clearly a sense in which, as a language, Java is more 
expressive than WSDL (though I would say that, in so far as WSDL *has* 
a semantics, it's prolly less expressive than OWL-S or OWL). Java is 
turing complete and has lots of practical expressivity (in terms of 
libraries, etc.). WSDL descriptions are, however, more abstract in that 
they define an *interface* which may be realized by code written in 
arbitrary languages. This adds a kind of expressivity!

> I am all but an expert but isn't this somewhat simplistic to use all 
> these translations transparently ?

Usually, there is some sort of augmentation or alteration done at each 
level. Well, java2wsdl is usually transparent *except* in so far as you 
might want to specify alternative bindings (which is significant!). 
When going form WSDL2OWL-S, you *at least* want to add to the profile 
something about the type, constraints, and capabilties of the service. 
So you've added informaton! Actually, it's similar to what you might 
add when creating a UDDI entry.

Hope this helps.

Cheers,
Bijan Parsia.
Received on Thursday, 2 December 2004 15:36:42 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Sunday, 16 March 2008 00:10:59 GMT