Re: [TED] Action-188, ISSUE: production rule systems have "difficulty" with recursive rules in RIF Core

> 
> > > > In particular, you didn't
> > > > address my attempts at defining what you call "conformance" more precisely
> > > > and also Frank's arguments.
> > > 
> > > Perhaps I missed them, although I think I read everything in this
> > > thread.  Pointer?
> > 
> > This was the main point of the message to which you were replying.
> > 
> > Anyway, here it goes again.
> > 
> > Conformance is to be defined by a semantics-preserving 1-1 mapping from a
> > rule system (RS) to a (possibly subset of a) RIF dialect (RIFD).
> 
> Can you rephrase that into a constraint on products?  That's where you
> lost me the first time through. 

I am not sure if "constraint" is the right term here, but whatever.

> 
> > To avoid a
> > misunderstanding, 1-1 means into, not onto. The mapping may only map a
> > subset of the RL for a whole number of reasons.
> 
> I'm okay with saying only a subset needs to be mapped -- every rule
> language is likely to have features that go beyond the nearest RIF
> Dialect -- but I think users need to be warned (at least in "standards
> mode") when they create rules outside the mapping.  Are you okay with
> that?

This is for the vendor to do. Vendors are the ones who define the mapping
for each particular product.

> > To exchange a ruleset, R, between RL1 (with mapping f1) and RL2 (with mapping
> > f2), you do inverse-of-f2(f1(R)). If f1(R) is not in the co-domains of
> > f2, you get an exception.
> > 
> > If the vendors of RL1 and RL2 have sufficiently good technical people, then
> > they should be able to describe precisely what the co-domains of f1 and f2,
> > then it would also be possible to describe the co-domain of f1 composed
> > with the inverse of f2. In this way you, as a user, would know what you can
> > or cannot exchange. But if those vendors don't have good technical people
> > then no big deal. Getting an exception is good enough.
> 
> I don't think getting an exception is good enough.

Vendors who can properly define their sublanguages will provide better
descriptions.

> I think users need
> to know when they buy a product which RIF dialects it fully implements.
> You'd agree, at least, that users would like that, right?   (Better to
> know the limitations at purchase time instead of run-time...)

Now you are talking about "implements" vs. "complies".
This is also what Bijan was aiming at, I think. See my response to him
http://lists.w3.org/Archives/Public/public-rif-wg/2006Dec/0098.html

I agree that there should be two notions here. Not sure which should be
called "compliance". I think the second one should be called "implements"
and not "complies". (Hope the quotes don't scare you this time around :-)


	--michael  


>    - Sandro

Received on Monday, 18 December 2006 19:08:04 UTC