[
Lists Home |
Date Index |
Thread Index
]
Paul Prescod wrote,
> Miles Sabin wrote:
> > Which implies giving up decidability.
>
> I do not believe that is true.
Believe what you like ... arithmetic is undecidable, so extending a DL
with arithmetic gets you an undecidable system.
Did you even read that paper? Or did you just grep it for something
which superficially sounded like it supported your argument? The
authors aren't talking about expressing general arithmetic contraints
in DLs, they're talking about using an arithmetic equation solver to
_implement_ standard DL inference (as opposed to using a tableau
system). That's what it says in the very first sentence you quoted!
You'd have been better off referring to, eg.,
http://citeseer.nj.nec.com/baader99expressive.html
esp. section 4. But that still won't show you how a constraint like
v=h*w*d can be expressed in a DL, because it can't be done.
Cheers,
Miles
|