Design
Here is a brief overview of the main data structure used to build the gcode from the 3D model. The original Skeinforge stores the results each plugin as text, initially SVG and then annotated gcode. This allows each plugin to act completely independently, but means each plugin has the overhead of parsing the text, and the approach is inherently serial. One benefit of storing the data in a shared structure is that it makes it easier to recode algorithms to make use of parallel processing. Another is that referencing specific parts of the model should be quicker.
The following diagram shows the main entities used to store the gcode representation of the model. Largely it is a collection of paths, and the translation to gcode commands is deferred to the very end of the process.
This diagram shows how the entities correlate with the actual output from the process.
