[gecode-users] bit-vector variables and constraints

[Christian Schulte]

Oh, then there is one more thing which I think you did not do: for bit
vectors support for modelling such as for integer variables (|, &, etc)
would be cool.

It can also be done by cut-and-paste but might be a bit harder.

Anyway, consider coming over.

Cheers
Christian

--
Christian Schulte, www.ict.kth.se/~cschulte/


> -----Original Message-----
> From: users-bounces at gecode.org [mailto:users-bounces at gecode.org] On
> Behalf Of Christian Schulte
> Sent: Thursday, October 23, 2014 11:25 AM
> To: kellen at cretin.net; users at gecode.org
> Subject: Re: [gecode-users] bit-vector variables and constraints
> 
> Dear Kellen,
> 
> Yes, I think that would be interesting. However, I word of warning: while
the
> stuff that remains to do is all easy, it might actually take some time to
get the
> stuff ready so that we could accept it as a contribution.
> 
> What remains to be done is (I am not 100% whether you might already have
> done part of it):
>  - Each file must be under the Gecode license. Easy: just add the standard
header
> with your name to each file.
>  - The structure of the contribution should be the same as for the other
variable
> types in Gecode (files, folders, file extensions, includes, namespaces, no
using
> clauses apart from individual functions, etc). Easy:
> just have a look at any of the variable types that come with Gecode.
>  - The set of branchers must be the same as for the other variable types
to get a
> uniform interface. Easy: just cut and paste will do (beware that you have
to
> retain the original licenses when you do stuff like this).
>  - The set of propagators must be sufficiently complete to be useful.
>  - We need some examples that use the variables.
>  - We need tests! Here again, the easiest is to start from the tests for
some
> other variable type in Gecode.
> 
> Some of the things sound like we are nitpickers. We are as we have to be:
> after all, when we accept a contribution we promise that we will try to
keep it
> alive (unless we find out later that nobody wants it).
> 
> The good news is, we are physically close. If you want to visit me in
Kista after
> you thought the stuff through and have a draft of your contribution you
are
> more than welcome. I'd even buy lunch ;-)
> 
> Cheers
> Christian
> 
> --
> Christian Schulte, www.ict.kth.se/~cschulte/
> 
> 
> > -----Original Message-----
> > From: users-bounces at gecode.org [mailto:users-bounces at gecode.org] On
> > Behalf Of kellen at cretin.net
> > Sent: Wednesday, October 22, 2014 10:30 PM
> > To: users at gecode.org
> > Subject: Re: [gecode-users] bit-vector variables and constraints
> >
> > Hi everbody!
> >
> > As Pierre wrote, the code available at:
> > https://github.com/kellen/bitvector-masters
> > is a result of my master's thesis in Computer Science with Pierre and
> Jean-Noël
> > Monette at Uppsala University.
> >
> > If there is interest, I would be willing to do the work to make the
> > code
> conform
> > with Gecode style/convention/optimizations. I am, however, a relative
> novice
> > C++ coder, as I have mostly been employed doing python and java.
> >
> > There are some obvious improvements to be made to the code, some of
> > which are detailed in my thesis, and probably other potential
> > improvements which
> will
> > be obvious to more seasoned C++/Gecode users.
> >
> > Christian/Guido: please let me know if/how I should go about
> > integrating
> the
> > variable implementation into Gecode core.
> >
> > Cheers,
> > Kellen
> >
> > _______________________________________________
> > Gecode users mailing list
> > users at gecode.org
> > https://www.gecode.org/mailman/listinfo/gecode-users
> 
> 
> _______________________________________________
> Gecode users mailing list
> users at gecode.org
> https://www.gecode.org/mailman/listinfo/gecode-users