Re: Handle graphics plotting functions

Petr Mikulik wrote:

> >> Unfortunately, the same calmness happened over my C++ 
> implementation of the
> >> histogram routine, which is more than 25x faster than the current 
> hist.m.
> >> See the thread "Re: fast hist() / histogram() implementation in 
> C++" here
> >>         http://www.octave.org/mailing-lists/help-octave/2004/
> >>
> > Writing octave core functions in C++ must be highly justified, and 
> speed is
> > not the only criteria, I myself have had an improved triu/tril 
> function in
> > C++ knocked back. Concretely the above thread showed upto 50 times 
> speed
> > improvement. But the current m-file on a test image took only 1 
> second. So
> > does even a 50 times speed improvement justify the issue of making code
> > maintaince much harder with the addition of C++ code..
>
> Try it on an image of 512x512, 1024x1024 or larger. It takes dozens of
> seconds by the .m code, and that's why the C++ improvement is a must.
>
You must convince John of this if the code is to go into octave itself. 
Its probbaly not a good idea to include a function with the same name 
and function as an existing function in octave, as the politic at the 
moment is to remove such things by migrating them back to octave or 
dropping them. Though inclusion in octave-forge under another name is 
possible.

> >> I would recommend that you contribute them to octave-forge,
> >> probably in the main/general directory.
> > Are you a member of octave-forge. "Us" I'm afraid includes "you", 
> and if you
> > want code included, you must take ownership of the code right to its 
> logical
> > conclusion of supporting its inclusion in octave-forge.
>
> No, I'm not on the list.
>
Do you want to be? If the answer is yes, then there is nothing easier 
than to make you a member of octave-forge with access to the CVS...

> I don't have and won't have time to maintain gnuplot subsystem in Octave,
> sorry. But that project would need someone to check/update the code for
> gnuplot 4.0 features, and it was what I have asked in my earlier mails.
> That person can get support from gnuplot developers from this or from
> gnuplot-beta mailing list.
>
Would you maintain hist, histc, and ginput? Maintaining your own code is 
all anyone can ask for..


D.

> ---
> Petr Mikulik
>


-- 
David Bateman                                [email protected]
Motorola Labs - Paris                        +33 1 69 35 48 04 (Ph) 
Parc Les Algorithmes, Commune de St Aubin    +33 6 72 01 06 33 (Mob) 
91193 Gif-Sur-Yvette FRANCE                  +33 1 69 35 77 01 (Fax) 

The information contained in this communication has been classified as: 

[x] General Business Information 
[ ] Motorola Internal Use Only 
[ ] Motorola Confidential Proprietary