Thanks, and yes: you answer your own question! Having a real interval solver
is indeed something for a 3rd party endevour.

However, let me try to make a point here. It is definitely not easy to get a
new constraint domain up and running in Mozart (I think that has become
clear from the communication here on the list). And, the interface comdemns
modules to be second-class citizens compared to finite domains and finite
sets: the interface does not allow to achieve the same expressiveness and in
particular not the same performance by far.

This is in strong contrast to Gecode. Gecode has no special pets: all
constraint domains including finite domains and finite sets use the very
same powerful interface. So, no penalty whatsoever. There is a contribution
CP(Graph) for graph variables. The authors have been able to get things up
and running (with substantial global constraints) just in a few weeks!

Sorry for the advertisement, but I think one of the biggest weaknesses of
Mozart is that the interfaces for constraint extension are not powerful and
by far not efficient enough! This has been one main motivation behind the
development of Gecode: have lean and powerful kernel that opens every
possible aspect for programming without sacrificing expressiveness and
efficiency.

Christian

--
Christian Schulte, http://www.imit.kth.se/~schulte/ 

-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On
Behalf Of Filip Konvicka
Sent: Thursday, January 19, 2006 9:09 AM
To: [EMAIL PROTECTED]
Subject: Re: Announcement: Gecode 1.0.0 released


Congratulations!

I've watched this for some time and I've been really looking forward to 
this. If there was a wish-list, I think that a real-number constraint 
module like XRI in Mozart would be great. (But this sure is suitable for 
a "3rd party" plugin.)

Having a C++ interface to all this is great - the thing I hate in Oz is 
that there's no typechecking...

Cheers,
Filip

____________________________________________________________________________
_____
mozart-users mailing list
[email protected]
http://www.mozart-oz.org/mailman/listinfo/mozart-users


_________________________________________________________________________________
mozart-users mailing list                               
[email protected]
http://www.mozart-oz.org/mailman/listinfo/mozart-users

Reply via email to