MAPM & GMT Package design?

[Clark Dunson]

Hey everybody;

I'm thinking of marrying GMT (Generic Mapping Tools) and Octave to create a 
mapping toolbox.  It's a sizeable project, so I started cutting my teeth by 
first bringing together MAPM (http://www.tc.umn.edu/~ringx004/mapm-main.html) 
and Octave as a learning exercise.  That was last week, and I got pretty far.  
I can auto-generate the bindings as oct-files with help info for DEFUN_DLD, 
made derived octave_base_value, etc.  But this morning I discovered cdemill's 
wrapping of GPM and the mptoolbox, and learned that we already have an APM 
toolbox!  Oh well, the learning experience was good, perhaps the biggest lesson 
being to check more thoroughly about whether I am duplicating effort! :)

So please let me check about mapping!

I like GMT, and have used it for years (license=GPL, 
http://gmt.soest.hawaii.edu/).  I feel that GMT has some advantages over 
MATLAB's toolbox.  One of the big disadvantages of the MATLAB mapping toolbox 
is its memory architecture.  All layers must exist loaded into MATLAB.  GMT on 
the other hand takes advantage of the ability to append Postscript files, thus 
each layer is a completely separate memory allocation.  This allows very large 
and beautiful print-quality maps to be made.  'nuff said, now questions:

Would I be duplicating effort, is it a useful thing for me to spend time on?  
Is anybody else on it or something similar, or does anyone want to help?

My thought is that Octave could improve my general usage of GMT by implementing 
a array/tree of map layers all derived from a base map layer object.  The 
command line interface to GMT that I use today is tedious.  But a command-line 
design for a toolbox, such that the toolbox can produce the equivalent command 
lines for easy regression could be very cool.  Prolly all this has been done 
already by someone else? 

blah blah blah

Anyways, Thanks for your time and any thoughts you might have!

Clark