Understanding the old approach

The previous version was based on bezier curves in Blender. You can create a curve and then apply a bevel to it using the shape of your choice.

• Issues
  • You can only apply 1 Object as bevel on all the curves on one layer. Which means you need a separate curve object for every feature on every layer because perimeters, infill lines and supports etc. will all have different line widths. This meant tens of thousands of layers even for the smallest of prints and Blender does not handle that many layers very well and it led to sluggish processing.
  • Applying a bevel to a curve creates a lot of edge cases, z-fighting, and other artifacts that always need workarounds.
  • All of it was procedurally generated which was not great for memory while rendering and led to slower renders
  • TL;DR Too much geometry and computation = slow Blender
• Advantages
  • You could control the visibility of every feature after loading the GCode, which is no longer possible
  • You could control the seam distance after loading the GCode which is also no longer possible

Understanding the updated approach

The updated version does away with treating GCode as curve objects and directly converts the GCode to mesh objects

• Advantages
  • No weird artifacts
  • We can fine-tune the texture of the end-points
  • z-fighting can be dealt with more effectively
  • Less geometry
  • Faster processing
  • Less number of layers
  • More accurate to the widths and heights in the actual gcode
  • TL;DR Faster processing = YAYYYY!!!!
• Disadvantages
  • We cannot interactively change the visibility of the features after loading the GCode. You can only pick which features you want before loading the GCode.