Re: ACTION-219: review of CORE (more)

> 
> 
> Hi again,
> 
> I want to add the following remarks to my review of the core document
> (which can be accessed at
> http://www.w3.org/2005/rules/wg/wiki/FrontPage?action=AttachFile&do=get
> &target=coreRev-AG.pdf)
> 
> This concerns the section titled "Intended Models of Rules."
> 
> I don't necessarily disagree with the details of that section (but see
> next paragraph), but I don't see what purpose it serves by being
> included at this point.

Quite a number of things here are in order to explain things to the group
members who don't have the background in this sort of things.
Jos' comments mention a bunch of such places.
There was a proposal (or decision?) to have an accompanying
architecture document, where I think such remarks (dialects, sorts,
semantic extensions, etc.) belong and should be put in a coherent framework
there.


> Also, there is a way of interpreting what goes on in a system that
> concludes "p" from "p :- naf q" that does not involve adopting a
> non-classical semantics.  Basically, one can view the naf operator as
> giving license to assume "not q" under certain circumstances.  Once
> that assumption is made, "not q" is deemed true and one may infer "p"
> validly from "p :- naf q" and "not q" according to classical semantics.

Is it really that simple? 

- What does it mean to "make an assumption"?
- What is the model theory of "making an assumption"?
- In what sense what you have in mind is more classical than, say, the
  stable model semantics?
- What if you have recursion through negation?



	--michael

Received on Monday, 12 February 2007 18:31:07 UTC