It is desirable for the control software to perform its basic functions of slicing/dicing, printing, and board testing without having to require a 3D card. The benefits are numerous:
* This would allow much cheaper machines to be used for the control PC.
* You could use a dedicated machine and possibly make it networked.
* Creating a command line version of the software would also be easier.
* It would also help eventual adoption in lesser developed countries.
I'm not familiar with the code, so some investigation is needed. my guess is that it would be good to abstract the build process a bit and create an object -> toolpath layer and a toolpath -> output layer.
we could switch to some sort of XML representation of the toolpaths required. we could either use and improve the rough format that Forrest and I devised, or investigate something like gcode. my guess is that going with our own format is the best route in the long term. we could always (and probably should) write an input plugin for gcode.
then, for the inputs we would be able to write code that would handle a variety of input files, and then produce the sliced files.
from there, the output would be able to do a variety of things like write to file, use a null device which would be for testing, a cartesian bot for actual printing, or even gcode for repstrapping via a pre-existing CNC machine with a thermoplast head.
there are probably alot of ways to go about this, and it is not top priority. it also means some major changes which are probably not appropriate when we are trying to get the thing out the door and working.