Re: Plotting status in Octave

[Thomas Weber]

On Tue, Oct 21, 2008 at 01:59:39PM -0600, Jonathan Stickel wrote:
> On 10/19/08 address@hidden wrote:
>> Date: Sat, 18 Oct 2008 14:40:46 -0500
>> From: " Jordi Guti?rrez Hermoso " <address@hidden>
>> Subject: Plotting status in Octave
>> To: "octave maintainers mailing list" <address@hidden>
>> Message-ID:
>>      <address@hidden>
>> Content-Type: text/plain; charset=UTF-8
>>
>> I seem to have lost track of what's been happening with plotting in
>> Octave. I know Shai Ayal and Michael Goffioul were doing a lot of
>> work[1] towards moving away from gnuplot, and I remember some efforts
>> to move their merges from octave-graphics into main. What happened to
>> this?
>>
>> And remember MathGL[2]? It looks very pretty. Is it going to be a part
>> of Octave?
>>
>> I am *very* interested in getting high-quality 3d plots into Octave.
>> In fact, nowadays my primary use of Octave is for post-processing and
>> producing nice-looking plots. Octaviz has produced the best-looking 3d
>> plots for me so far, but it's abandoned upstream and I now agree with
>> Thomas Weber's opinion that combining VTK and Octave is too much work
>> and too resource-intensive. Compiling Octaviz requires compiling all
>> of VTK with some flex and bison arcana that I am unable to debug.
>>
>
> As much as I love VTK/Octaviz, I have recently reached the same  
> conclusion.  I would gladly continue to develop the frontend (m-file  
> functions) of Octaviz, but with Dragan gone and no other c++ guru to  
> debug the internals of Octaviz, I am afraid it is dying (or already  
> dead).  I guess the best way to use octave and VTK together in the long  
> term is to save data from octave and write VTK scripts/programs to  
> generate the plots using the saved data, as I have done in the past.

I actually disagree with both of you. The current way of Octaviz is
dead, yes. But, what is Octaviz currently? It provides bindings to VTK
from Octave. That means that 
a) We have far more functionality then we ever need (like special
functions for long unsigned ints, or something like that).
b) Building it takes ages. 
c) Debugging it is extremely difficult.

What it could develop into is a different backend to Octave, by only
providing the needed functionality (and Octave would link to VTK,
probably by one .oct file). 

This is a long-term project, but hey, maybe someone has lots of time.

        Thomas