Re: FOL versus Rule Languages - A tutorial

jos.deroo@agfa.com wrote:
> [...]
> 
> > In NAF, the scope is implicit, but well-defined. However,
> > this implicitness doesn't work well on the web because
> > the Web is practically infinite.
> 
> true
> 
> > There is always that "somewhere on the web" as you put it,
> > which the engine may not be aware of. One way out of this
> > is to let the user specify the scope of the inference
> > explicitly (which is what SNAF really is). In this way,
> > you tell the engine where to look.
> 
> Right, but then you say, and I now understand that SNAF
> is non-monotonic; I'm fine when such reasoning results
> are used for local action (and have good experience with
> that) but I really don't see it for cases where we have
> merging of rulesets created without knowledge of what
> they would be merged with (to use Sandro's words)

Jos,

I didn't say that nonmonotonic reasoning is the only game in town for
everything. Merging rules is a whole different ball of hair. As far as I
can tell, logical reasoning (monotonic or nonmonotonic) is not what you
would use here. This is an intellectual activity, which can be automated
with the help of heuristics (which are not logical inference rules),
machine learning, etc.

I bet, however, that **if** you decide to encode those rule-merging
heuristics in a rule-based language (and not in Java, for instance) then
you would prefer a Prolog-like language with S/NAF rather than FOL.


	--michael

Received on Friday, 26 August 2005 20:28:38 UTC