RE: Relation between OWL and OWL-S

Bijan,

Hope CMU's CODE and your efforts in HTN planning in SHOP2 make OWL-S more
popular. Easy-to-deploy tools are so important.

Cheers

Jun

-----Original Message-----
From: Bijan Parsia [mailto:bparsia@isr.umd.edu] 
Sent: Wednesday, November 24, 2004 12:05 AM
To: jshen@it.swin.edu.au
Cc: public-sws-ig@w3.org; daniela.claro@eseo.fr
Subject: Re: Relation between OWL and OWL-S


On Nov 23, 2004, at 6:20 PM, jshen@it.swin.edu.au wrote:

>
> Hi, Dan,
>
> I have no answer to your question. But from my understanding and 
> practice, I
> think we need not worry about difference or relations between OWL and 
> OWL-S.

That may be true, although there are interesting expressive gaps 
between the intended meaning of OWL-S descriptions (esp. in the process 
model) and what is actually said by the OWL which encode them.

> You can decide how to describe what you need, no matter using RDF, OWL 
> or OWL-S
> or even RuleML.

Eh. I recommend OWL-S, naturally :) Or, rather, unless you are doing 
something entirely experimental in service description frameworks, 
trying to work with one of the well developed ones out there. Reuse, 
y'know.

> Esp. OWL-S and RuleML are not appearing as a mature
> recommendation,

OWL-S is now a member submission to the W3C.
	http://www.w3.org/Submission/2004/07/

It's fairly mature for some crufty sense of mature.

> by anybody. We are not expecting every human or program or
> agent become able to understand between each other yet.

This seems like a non sequitur.

>  The important thing is
> how to make your work done, then is how to present them. In the latter 
> case,
> more intelligent people or agent make more complex or smarter toys or 
> tools or
> applications or killer systems, like google (froogle, :), so that you 
> need a
> good language to express it, including typing relations, let's say.

Er...more into the weeds for me.

Cheers,
Bijan Parsia.

Received on Tuesday, 23 November 2004 23:53:54 UTC