First-class constraints

[Alberto Delgado]

Hi Harmon,  you can find the tobias muller's contribs inside 4 directories,
fcp,  reflect,  investigator and davinci.  The implementation of the module
is just a little bit different from the paper.  You'll get a good example
from the fcp directory,  the name of the file is opi.oz

if you want to compile the contribs with the source code you can use the
configure like this:

./configure --prefix=/your_mozart_path --enable-contrib --enable-contrib-fcp
--enable-contrib-reflect --enable-contrib-davinci
--enable-contrib-investigator

after the make and make install command you will find the contribs' functors
on the /cache/x-oz/contrib directory of your mozart directory

A guy here implemented a couple of examples that use the fcp module,  i
could send then to you if the opi.oz example isn't enough.

Check the contribs out and if you need something more general i could send
your our  implementation.

Att
Alberto

An remember,  some functions don't work with mozart 1.3.1

On 12/2/05, Harmon Nine <hnine at isis.vanderbilt.edu> wrote:
>
>  Hi, Alberto.
>
>
>
> I've looked at the "mozart-1.3.1.20040717/contrib" source directory, but
> wasn't able to find the "constraint" package as described in Muller's
> "Promoting Constraints to First-Class Status" paper.  Is it under some
> directory other than one named "constraint"?
>
>
>
> Here's a listing of the directory "mozart-1.3.1.20040717/contrib" source
> directory:
>
> Makefile.in  configure.in  dpanel  investigator  maple     reflect  text
>
> ap           davinci       fcp     io            micq      regex    tk
>
> compat       directory     gdbm    lp            os        ri
>
> configure    doc           gtk     makefile.oz   psql      ri_lp
>
>
>
> Maybe one of these directories contains the package (maybe it's been
> removed?)  I've looked at the code in them, but none of them looked like
> they were it.
>
>
>
> If it's not there anymore, do you know a place I might obtain it?  Or, if
> your mods are available, could we try using them?
>
>
>
> Thanks.
>
> -- Harmon
>
>
>  ------------------------------
>
> *From:* Alberto Delgado [mailto:trosky99 at gmail.com]
> *Sent:* Friday, December 02, 2005 12:33 PM
> *To:* Harmon Nine
> *Cc:* mozart-hackers at ps.uni-sb.de
> *Subject:* Re: First-class constraints
>
>
>
> Hi, i don't know anything about the Mozart's Debian package,  i know you
> can get it  either from the tarballs of the mozart's website or  from the
> mozartCVS.  you'll find it inside the contribs' directory,  it is inside
> mozart's directory. The fcp and reflect directories have the important
> things to use the first-class module,  you can also play with the davinci
> and investigator contribs.
> We think the idea of the module is very  good,  but it is a prototype,
> so  if you want to work with that,  it'll need a lot of work.
> I remember that all the module works very well on mozart 1.2.5,  but some
> functions have problems on mozart 1.3.0.
>
> We developed a modification to the mozart's source code that allow you to
> stop one or more constraints,  put then together and other stuff.  we used
> the ideas and a lot of information from the First-class module. The idea of
> our  modification is something  like http://
> discipl.inria.fr/pub/milestone2/D.WP3.1.M2.1.pdf ,  without the graphical
> interface.
> We also worked   with the reflect module.
>
> Try with the contribs and if you have any doubt i'll try to help you.
>
> Att
> Alberto Delgado
>
>
>
>  On 12/2/05, *Harmon Nine* <hnine at isis.vanderbilt.edu> wrote:
>
> I can't seem to find it.  I'm using the mozart and mozart-stdlib packages
> on the Debian distribution.  There are packages like 'regex' and 'gdbm'
> under the contrib directory, but no 'Constraint' package.  Is it something I
> have to download separately?  I can't find it on Mogul, either. L   Is
> there a particular site where it is available?
>
>
>
> We'd like to use a first-class constraint system for scheduling.  The
> scheduling tasks we're tackling are very complex, and sometimes run
> "forever," never returning a result.  It would be invaluable to be able to
> tally, at some point in the search, which constraints are causing the most
> failures.  From there we could perhaps turn them off, or provide
> modifications to the environment so they are more likely to be satisfied.
> I'm sure first-class constraints could help in other ways as well. J
>
>
>
> Thanks very much for any help you can provide.
>
>
>
> -- Harmon
>  ------------------------------
>
> *From:* Alberto Delgado [mailto:trosky99 at gmail.com]
> *Sent:* Thursday, December 01, 2005 6:04 PM
> *To:* Harmon Nine
> *Subject:* Re: First-class constraints
>
>
>
> Hi,  you can get it from the mozart's contrib directory,  it's not a
> modification to the oz code, it works like a functor.
> What do you want to do with it?
> I did some work with that in the Avispa Research group.
> We modificated the oz code to do something similar to the mullers'
> first-class constraints
>
> On 12/1/05, *Harmon Nine* < hnine at isis.vanderbilt.edu> wrote:
>
> Does anyone know where I can obtain Tobias Muller's modifications to the
> Oz code that allows first-class constraints, i.e. via the "Constraint"
> module?
>
>
>
> Any help is greatly appreciated.  Thanks.
>
>
>
> -- Harmon
>
>
>
> _________________________________________________________________________________
> mozart-hackers mailing list
> mozart-hackers at ps.uni-sb.de
> http://www.mozart-oz.org/mailman/listinfo/mozart-hackers
>
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.gforge.info.ucl.ac.be/pipermail/mozart-hackers/attachments/20051202/c6459105/attachment.html