[Fwd: The LP Module (Again)]

[Paul Snively]

Gentlemen,

Thanks for your attention to my question. You're both correct: my  
patch is for the LP module found at gna.org for the simple reason  
that the one from Oz's "contrib" directory isn't installed by the  
Mozart-Oz 1.3.2 installer for Mac OS X, and in googling for it I  
found the gna.org one. :-)

My patch allows the module to compile and run on all of the examples  
given in the docs on the gna.org site; I'm quite confident that the  
patch is "correct." However, in the custom constraint propagator  
tutorial on the Mozart-Oz.org site, there's an intriguing example  
demonstrating combined FD constraints and LP by means of RI.intBounds  
that I have been unsuccessful in getting to work, but I'm sure that's  
due to transcription errors rather than bugs. Nevertheless, I'm  
stumped at present, so any insights that anyone can offer as to how  
I'm in error would be greatly appreciated.

Best regards,
Paul Snively

On Feb 26, 2007, at 6:09 AM, Gustavo Gutierrez wrote:

> On 2/26/07, Boriss Mejias <boris.mejias at uclouvain.be> wrote:
>> Dear Hackers,
>>
>> Here is a proposal of Paul Snively for fixing some bugs in the LP
>> module. In order to take a decision about this, I would like to
>> understand some things.
>>
>> Are these fixes meant for the LP code included in mozart/contrib/ 
>> lp ? In
>> that case, I'm confused because the patch doesn't match the source  
>> code.
>>
> No there aren't.
>
>> Is there any other LP module?
>>
>> Is there anybody with enough knowledge about the LP module to give an
>> opinion about the patch?
>
>
> Some time ago i reimplemented  the real interval module because the
> one made by Tobias Mueller had some problems and did not provide
> enough operations to suit our needs. I build the new module from
> scratch and uploaded it to gna.org.  This new module never has been
> part of the  official mozart repository. However i thin it offers
> several advantage over the official one.
>
> As the LP module provided in the mozart sources relies in the real
> interval module i also implemented a new module for that. Both modules
> can be found at the https://gna.org site.
>
> I think there is an important question: Should we include the new ri
> and lp modules in the mozart distribution?.  Or maybe we can put them
> in the mogul repository.
>
> Cheers,
> -- 
> Gustavo Gutierrez