W3C home > Mailing lists > Public > public-rule-workshop-discuss@w3.org > June 2005

Re: NAF v. SNAF - where is this being addressed?

From: Ed Barkmeyer <edbark@nist.gov>
Date: Wed, 29 Jun 2005 16:37:05 -0400
Message-ID: <42C30671.9080908@nist.gov>
To: Jim Hendler <hendler@cs.umd.edu>
CC: public-rule-workshop-discuss@w3.org

Jim,

I understand that your particular concern is to point to SNAF as a 
"required concept" in the solution to the NAF problem, and I fully agree 
with that.

But we seem to have opened another can of worms here.

you wrote:

>  More specifically, I think some of your points below related to what I 
> called in a different thread the difference between a "Web Rules 
> Langauge" and a "Semantic Web Rules Language" -- I've argued these 
> aren't identical.  If what we want is a syntax for the exchange of rule 
> sets (i.e. I can write a set, you can read them and use them as they 
> are) then IMO we should build an XML schema for rules and have a very 
> simple semantics underlying it (so as not to overconstrain solutions).  

I agree.  The point I made in the final Panel for the Workshop is that 
this part is well underway (I wouldn't say 'done') in the RuleML effort 
and the OMG Production Rules Representation effort (which are somehow 
linked).  I don't know that the concept "very simple semantics 
underlying it" is viable, and I am convinced that both of those efforts 
need some help in defining their semantics.  But I defer to Ben Grosof 
and the SWRL folks in that area.  So those standards have their place, 
and they may need work, but W3C does not (ostensibly) need to go there.

What we do need to do is to deal with three other concerns:
  - the relationship of rules to OWL
  - the relationship of rules to RDF
  - the distributed/linked ruleset

> On the other hand, if the goal is a distributed set of rules with the 
> ability to link to other sets, reuse parts of other business processes, 
> cut and paste one set of rules and edit for another application, etc. 
> (which, as reflected in the OWL FAQ [1] was the hard part of going from 
> traditional KR to OWL) then the task is harder, and NAF/SNAF in 
> particular becomes trickier to apply and we need to be careful about how 
> we delineate the goals.

So we agree that addressing the issue of distributed rulesets is a W3C 
concern that goes beyond the (planned) work of the other groups.

The OWL FAQ to which you pointed contains this observation about KR for 
the Semantic Web:

> [OWL] uses the linking provided by RDF to add the following
> capabilities to ontologies:
> 
>   * Ability to be distributed across many systems
>   * Scalable to Web needs
>   * Compatible with Web standards for accessibility and
> internationalization.
>   * Open and extensible

[Scalable to Web needs is a ringer.  The interpreting engine has to 
access, and potentially use every element of, the distributed resource, 
and scalability is about the limits of that engine.  The W3C role is 
about distribution, access standards, and extensibility.]

Yes, we can and should do this for Rules.  It is a useful enabler.

What it enables may be analogous to giving a child an automatic weapon, 
but that is the subject of a different diatribe. ;-)

>  btw, I think the jury is still out as to whether a Web Rules Language 
> or a Semantic Web Rules Language is preferable and to be chartered - but 
> if it is the latter, we have to resolve what we're trying to do more 
> carefully.

I would be happy to leave the Web Rules Language -- the exchange of a 
closed body of rules -- to the RuleML and PRR folk, but I think W3C 
would have to build on that to do the distributed and extensible part. 
So it would be good to work together from this point on. Further I think 
we really need to talk about the relationship to OWL ontologies and 
perhaps to ISO CL.  (I would like to avoid the UML/OCL mistake of having 
unconnected semantic models for two editions.)

FTR, I think the jury is still out on whether OWL is really a Semantic 
Web Language for distributed, extensible ontologies or just a lingua 
franca ("Web language") for DL+ ontologies.  So far, all I see is the 
latter, but OWL is young, and you have to have respected reference 
ontologies before anyone else can use them.

-Ed

-- 
Edward J. Barkmeyer                        Email: edbark@nist.gov
National Institute of Standards & Technology
Manufacturing Systems Integration Division
100 Bureau Drive, Stop 8264                Tel: +1 301-975-3528
Gaithersburg, MD 20899-8264                FAX: +1 301-975-4694

"The opinions expressed above do not reflect consensus of NIST,
  and have not been reviewed by any Government authority."
Received on Wednesday, 29 June 2005 20:37:10 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:16:21 GMT