Re: Draft prooosed 4th goal: RIF should be usable as the basis for a semantic web rule language

ewallace@cme.nist.gov wrote:

> Dave Reynolds wrote:
> 
> 
>>To discharge my first action from f2f3 I have drafted an initial write 
>>up my proposed fourth goal. It is on the Wiki at:
>>
>>http://www.w3.org/2005/rules/wg/wiki/RIF_should_be_usable_as_the_basis_for_a_semantic_web_rule_language
>>
>>and is linked from the page or recently added open issues.
> 
> 
> You can add my name to the list of champions for this goal.  Achieving 
> this goal was my motivation for advocated the creation of a Rule WG in the 
> Semantic Web Activity, and why I joined RIF.  

Thanks, done.

> A couple of comments:
> 
> From the text of the proposed Goal:
> 
>>To achieve interoperability we need one, or some *small* number of, recommended 
>>semantics that we believe is compatible with, and useful for, the semantic web.
>>
>>By saying 'basis for ...' then we allow RIF to, for example, omit a human usable 
>>syntax. We also allow the possibility that the recommended profile is an 
>>informative rather than normative part of the spec. For example, it might be a 
>>suggested RIF profile defined in the "Recommendation on using this rule 
>>interchange format in combination with OWL" deliverable.
> 
> 
> 1. I don't follow how this would allow RIF to omit a human usable syntax.

In the f2f discussions there was concern that RIF is an interchanage 
format and there might be other things needed to make a dialect usable 
as a language in its own right that the WG group wouldn't want to do. 
That the XML syntax will be horrible for humans was one comment. I agree 
that there are separate use cases which call for a human usable syntax 
so this may be a null issue.

However, *if* it turned out that RIF didn't have a bearable syntax I 
could live with it from the point of view of this goal. E.g. a subgroup 
could suggest a non-normative syntax for the semantic web rule subset 
perhaps as a W3C Note outside of the RIF rec-track documents and we 
would still have made progresss.

[It's also one attempt to give some semantics to the weasel-phrase 
"basis for". My assessment was that without that phrase this goal would 
not be acceptable to the WG. By putting this in we open the possibility 
that we can find a meaning for "basis for" which makes it acceptable to 
all sides. ]

> 2. I don't read the above comments as putting this proposed goal in opposition
> with the requirement for multiple formal semantics (and thus the wide coverage 
> FCS).  This is because the profile for SW can select a small set of 
> recommended semantics, even if RIF supports more.

Agreed. It just requires that the range of semantics cover the ones that 
are useful for this purpose and that some part of the WG put effort into 
selecting the small set.

Dave

Received on Friday, 9 June 2006 06:43:33 UTC