Re: 3D plotting solutions

[Johan Jansson]

Hi Ole,

On Fri, Dec 05, 2003 at 04:46:10PM +0100, Ole Jacob Hagen wrote:
> Hi, Johan.
> 
> I am developing Oplot++, which will support handle graphics routines of
> Matlab in Octave. I'm trying to make a best solution in design, since a
> lot of people want to use handle graphics functionality inside Octave. 
> This requires that a low-level handle graphics functionality is inside
> Octave, and sent to the visualisation application. 

Yes, I realize it's necessary to have such a system to be able to
refer to and manipulate multiple figures.

> I'm very excited to see that you have made pde-plot using
> Coin3D/OpenInventor. ;-) 
> Do you want to help me with pde-plot parts inside Oplot++? 
> This would be great, if you have opportunity and time to help us with
> Oplot++, with this matter. 

Sure, I'd be glad to help out with constructing a rudimentary pdeplot
for Octave/Oplot++. We should do this in such a way so that it can
work with other visualization applications as well.

I looked a bit at Oplot++ before, I will try to get it working so I
can do some testing.

> 
> Oplot++ is based on Coin3D and Qt. I'm a patriot, since I'm Norwegian,
> and both Coin and Qt is developed in Norway....;-) 
> Coin is used for visualisation, while Qt is for gui-widgets. 
> Non-gui signal/Slot mechanisms should be changed to use libsigc++,
> instead of Qt's signal/slot mechanism, though. 
> 
> Coin3D is released under GPL-license. 
> 
> I'm currently away until January, but I will check this email
> frequently. 
> 
> Hope you will consider to become a Oplot-developer. I believe you have
> a lot to contribute. 
> 

Yes, I hope so too. I would like to keep things simple
though. Concretely, the most important thing is to be able to
visualize the data you're working on, so you can see that it's
reasonable. It's less crucial if you can make pretty plots or plot
several plots on one page, or plot very efficiently etc. Currently
it's not possible to plot these things at all, so any ability is
infinitely better than that.

I also don't think it's essential to try to get 100% compatibility
with Matlab. I haven't worked that long with Matlab, but it seems as
if the feature list is growing uncontrollably. As long as the
essential features are there, and perhaps the nonessentials are just
empty implementations (to ensure source compatibility), then I think
users will be satisfied.

  Johan