[m-dev.] Re: lp_solve distribution

[Tyson Richard DOWD]

Fergus Henderson wrote:

> [A letter to Michel Berkelaar]

Thanks for doing that.

I've been talking with Tom, who has implemented a LP solver in
Mercury. Chances are we could use this code, although there are a number
of issues involved in actually integrating it with the compiler.
For completeness, the issues are:
	1. Numerical stability of the solver (probably not a problem
	   with our datasets).
	2. It should be a separate library, so there will be some work
	   involved in integrating it with the compiler (probably easier
	   than integrating clpr).
	3. It uses a different interface, so the termination analysis
	   code would need to be modified to use this interface.
	Did I forget anything?

Certainly it would be easier to use lp_solve in the short term
at least. 

-- 
       Tyson Dowd           # To fix this, edit BLAH\BlahKey\Blah\Whatever 
                            # in the registry.
     trd at cs.mu.oz.au        # *WARNING* Editing the registry can DESTROY
http://www.cs.mu.oz.au/~trd # your machine forever. Do not do it.