Re: Matlab central and octave

[Dmitrey]

Hi David,

David Bateman пишет:
> However, the problem is there are lots of people who
> are willing to suggest what needs to be done and few to actually do the
> work. If there was someone committed enough to support such a forum,
> then I'd support them wholehearted, but till someone steps up and says
> something like, "I've just created a test website implementing MFA for
> Octave, and will support it in the future" then I won't believe in it..
>   
Maybe, after this issue someone will apper?
As for me, I would gladly commit my OpenOpt solvers to the area - even 
now some of them are good enough, for example ralg() on non-smooth & 
noisy tasks is better than MATLAB fminsearch (/fminsearchbnd, 
fminsearchcon, fminsearchOS), adsmax 
<http://octave.sourceforge.net/index/f/adsmax.html>, mdsmax 
<http://octave.sourceforge.net/index/f/mdsmax.html>, other nelder-mead 
<http://octave.sourceforge.net/index/f/nelder_mead_min.html> 
implementations, even commercial snopt(), at least at 95% cases that I 
have tried (but I can't say I tried too much*). *Now it can handle all 
type of constraints - lb, ub, Ax<=b, Aeq*x=beq, c(x)<0, h(x)=0. New 
feature - nonSmoothSolve - fsolve analog for non-smooth funcs. I intend 
to publish new relize 5-10 days later, currently you can taste latest v 
0.35 at box.net/public/6bsuq765t4

**
> All of the people currently developing code for octave and octave-forge
> are all to busy on other projects, so it would take someone new to step
> forward to do this work..
>
> D.
I don't know anything of the situation, so I can only assume that you 
are right. However, I have no info what are you doing now (since Octave 
(like it seems to me) doesn't inform their users what will next reliase 
contain).
As for me, I guess, 2 key features that Octave miss vs MATLAB or even 
free Python(/scipy, numpy) is
1) at least one convinient IDE (containing easy settlement of 
breakpoints, better - conditional BP), for example eclipse wrapper - 
some years ago, as I noticed, there was a discussion about eclipse but 
no further steps were made - and I guess it was great error.
2) JIT

Best regards, D.