Re: Information-Theory functions.

On Sat, 2006-09-30 at 00:33 +0200, David Bateman wrote:
> Muthiah Annamalai wrote:
> 
> > While I work on directions you have suggested (Thanks!)
> > I did a preliminary web-search on existing matlab functions,
> > and cant seem to find anything like what we are looking at.
> > 
> > So might infact turn out to be useful, or atleast so I hope.
> > 
> > Also the info-theory.m is there just for the unoptimized version
> > of the functions, which dont need to be distributed, just for
> > reference implementations.
> > 
> > Thanks again,
> > -Muthu
> 
> Don't worry, if its useful matlab will implement it with an incompatible
> interface. And Muthu, I don't think either Paul or I decide in anyway
> what should go in octave-forge. Its supposed to be easily accessible for
> development and so package quality is expected to be variable.
> 
> Please read the file www/new_developer.html in the CVS after doing
> 
> ./configure; make; cd www; make
> 
> on the CVS for instructions of how to write to the new package manager
> specifications in octave-forge. I hope this will get to the website
> soon, but that's up to Soren, or me (someone else?) getting the time to
> finalize some of the issues (building proper from the built INDEX files
> (fixed?), linking correctly to the download page for the individual
> packages, including any additional documentation for a package in the
> package webpage)
> 
> Also, I think the best way to handle the importation of code into
> octave-forge is to let you do it yourself. Then the support issue is
> yours :-) .... You don't have an octave-forge account as far as I can
> see. What is your source-forge ID so I can add it?
> 
> Cheers
> David

David,
My source forge id is 'gnumuthu'. 
Im in the process of polishing whatever functions I have in hand
from earlier work, and adding tests to it. So I must be doing it
soon.

> Don't worry, if its useful matlab will implement it with an incompatible
> interface.
like the controls toolbox story?

Cheers
Muthu