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

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

From: Dave Reynolds <der@hplb.hpl.hp.com>
Date: Thu, 30 Jun 2005 14:17:04 +0100
Message-ID: <42C3F0D0.8020309@hplb.hpl.hp.com>
To: public-rule-workshop-discuss@w3.org

Ed Barkmeyer wrote:

> 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

Agreed that those are three concerns that could distinguish a W3C rules 
effort from the existing efforts. However, an effort which focused on a 
subset of those might still be useful and sufficient for a first pass.

For example, we have a particular interest in the use of rules for RDF 
processing. We would like such rules to have a well-defined relationship 
with OWL (but to not require all conformant RDF rule processors to also be 
complete OWL processors). However, we would be happy with distribution in 
the weak sense of "publish for reuse" rather than necessarily in the strong 
sense of "linked rulesets for general web-scale KR".

Example: an RDF processing agent wants to be able validate that a given 
package of data is sufficiently complete to be processed further and would 
like to advertise the integrity constraints it applies. Rules would be one 
quite useful way to do that.

Example: a data transformation agent takes a collection of data sets in RDF 
form and returns a processed abstraction/transformation/integration of that 
data. It wants to publish/load the nature of the relationship between the 
transformed data and the original data. A standard rule language would be 
one convenient way of doing that (though might require closed world 
operations like counting and non-datalog operations like object introduction).

In these cases the agent (this is agent with a small "a") is defining the 
scope of the application of the rules rather than the rules themselves.

Whilst I wouldn't want to push the analogy too far it seems to me that XSLT 
plays a somewhat analogous role in the XML stack. It is true that XSLT 
scripts can be linked but that seems to be used more for modularization 
than for distributed web-scale transforms. They tend to be isolated things, 
applied locally. That doesn't stop XSLT being a useful, commercially 
significant, web-related standard that it was entirely appropriate for W3C 
to have developed.

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

A concern, not the only concern.

Dave
Received on Thursday, 30 June 2005 13:18:00 GMT

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