Re: [Plplot-general] Comments and suggestions for PLplot 5.1.0

[Roberto H. <ro...@ad...>]

Geoffrey Furnish wrote:
> Roberto Hernandez writes:
>  > 3. One of the items mentioned in the "Remaining Important Bugs" section 
>  > of the new version announcement has to do with the 3D shading. While 
>  > searching for more information on the subject I came accross Generic 
>  > Mapping Tools (http://gmt.soest.hawaii.edu/gmt.html). It's a GPL 
>  > application which produces shaded plots (among other types) from data 
>  > sets and outputs EPS files. The website includes some very impressive 
>  > examples. Maybe the 3D shading portion of the code could be adapted to 
>  > PLplot.
> 
> We can't pull GPL code directly into the LGPL PLplot without losing
> the leading L.  For things like this, we really need to start with
> mathematics, and code from there.  Again, this is a project where
> there is no one currently signed up, I think, so submissions would be
> welcome.   (Of course looking at what other packages have is fine, I
> think, but we just can't directly leverage GPL code without a license
> brouhaha). 

I've been doing some thinking about this issue and I came up with a 
question.

The 2D "shaded" plots are great as they are now. More specifically, I 
use PLplot under Octave and the most comfortable representation of R2 
--> R functions for me is the "shade" command. Wouldn't it be a good 
idea to have a combination of the mesh and shade type plots?

Suppose I have a function f(x,y). Creating a surface plot of that 
function and coloring it according to the position of a light source is 
complicated. Maybe just having a mesh plot, where the lines are 
color-coded according to the value of f(x,y) would be a lot simpler. 
Especially considering that the algorithm for the mesh is taken care of 
and so is the color coding of f(x,y). Sometimes these colored mesh plots 
are even more useful than surface plots (I would imagine most of you 
have seen these types of plots in matlab).

I haven't looked at the code and I don't know whether this would really 
be easier to do, but it seems reasonable. Just thought it might be a 
useful suggestion.

Regards,
Roberto