lp_solve replacement (future project)

["Pedro F. Giffuni" <gi...@tutopia.com>]

Hi again;

It's not at all urgent, but if someone has an inclination for
a nice Linear Programming solver, I would point to COIN-OR's
CLP Project:

http://www.coin-or.org/projects/Clp.xml

cheers,

Pedro.

Re: lp_solve replacement (future project)

["Pedro F. Giffuni" <gi...@tutopia.com>]

--- On Fri, 6/24/11, Mathias Bauer <Ma...@gmx.net> wrote:

> From: Mathias Bauer <Ma...@gmx.net>
> Subject: Re: lp_solve replacement (future project)
> To: ooo-dev@incubator.apache.org
> Date: Friday, June 24, 2011, 2:22 AM
> Hi Pedro,

> let me quote Niklas Nebel from the OOo dev list:
> 
> > Dependencies on LGPL libraries may soon become a
> problem. One such
> > library is lp_solve, which is the basis of the
> OpenOffice.org Linear
> > Solver in Calc.
> 
(snip)

> http://qa.openoffice.org/issues/show_bug.cgi?id=118160
> 

Extremely interesting, and another good reason for
bringing in the bugzilla database ;-)

I will get in touch with Niklas as I have some
personal interest in this.

Thanks!

Pedro.

Re: lp_solve replacement (future project)

[Mathias Bauer <Ma...@gmx.net>]

Hi Pedro,

On 24.06.2011 03:45, Pedro F. Giffuni wrote:
> Hi again;
>
> It's not at all urgent, but if someone has an inclination for a nice
> Linear Programming solver, I would point to COIN-OR's CLP Project:
>
> http://www.coin-or.org/projects/Clp.xml
>
> cheers,
>
> Pedro.
>
let me quote Niklas Nebel from the OOo dev list:

> Dependencies on LGPL libraries may soon become a problem. One such
> library is lp_solve, which is the basis of the OpenOffice.org Linear
> Solver in Calc.

(snip)

> I created issue 118160 and attached a simple proof-of-concept patch
> to use CoinMP instead of lp_solve. From some limited testing, this
> seems to work very well. Some changes will still be needed for
> building and packaging, and perhaps some tweaking of solver settings.
> But the patch shows that replacing lp_solve is clearly a (pardon the
> pun) solvable problem.
http://qa.openoffice.org/issues/show_bug.cgi?id=118160

Regards,
Mathias