"16.1 Two-Dimensional Plotting" is unclear

Søren Hauberg soren at hauberg.org
Sat Dec 15 07:08:57 CST 2007


lør, 15 12 2007 kl. 04:57 -0800, skrev Sergei Steshenko:
> Thanks for the 'figure' tip.
> 
> "help figure" in octave-2.9.19 does return some info, but
> 
> http://www.gnu.org/software/octave/doc/interpreter/Two_002dDimensional-Plotting.html#Two_002dDimensional-Plotting
> 
> page has nothing on 'figure'.
But this page does:
http://www.gnu.org/software/octave/doc/interpreter/Multiple-Plot-Windows.html#Multiple-Plot-Windows

(I just followed the link that said "Multiple Plot Windows" which I find fairly descriptive to your problem)

> I am not sure I understand octave developers' approach to plotting.
> 
> For example, wxWidets developers decides (if I understand them correctly):
> 
> 1) here is our set of classes, i.e. data + methods;
> 2) here is our underlying code implementing the above set on top of native
> toolkit (like gtk+ for UNIX-like systems);
I haven't contributed to the plotting system, so I don't know the
details. However, comparing the Octave plotting system to wxWidgets
doesn't IMHO make any sense at all. wxWidgets is a general system for
creating graphical user interfaces, while Octave provides functions that
allow you to visualise some data. Two very different objectives.

> So, what do 'octave' developers intend to do ? To replicate whatever Matlab does
> (that's more or less like the 1) above) by mapping it on GNUPLOT (that's more or less
> like the 2) above).
Compatibility with matlab is important. Like it or not, currently matlab
is one of the most used tools in many fields of science. There is a lot
of matlab code out there, so if you want to use the code that other
people developed you need something that is compatible with matlab. So,
from what I understand, the current strategy is to copy whatever matlab
does and map that to gnuplot. In the future gnuplot might be dropped,
but the compatibility goal will likely remain.

On a personal level, I'm not a fan of the handle graphics API that
mathworks has created, but I don't think we are in a position to not
implement it (yet).

Søren



More information about the Bug-octave mailing list