Well, yes it looks stable but here are some topics I'd like to address:
- I still have to figure out how to apply presets. The original code from Enrico can't work anymore with 2.56 because it's no more possible to alter setting/parameters from the drawing part of the interface. I don't know if it's a change made by the Blender team or a bug within the new API. As callbacks seems not supported either at the moment I'm still looking for a workaround.
- user-defined presets: if I can fix the previous one I'll investigate here. No emergency here as it would anyway not allow users to save significant time in their workflow. Just nice to have.
- more fool-proof checking to warn user at an early stage something is wrong (such as bad texture names or non-existing paths)
- code cleanup: I'm trying to split out as many functions as I can to make the code more readable and easier to maintain. It should also make new command line options of Octane easier to implement (if any).
- Camera animation: I should be able to quickly implement this one as soon as I'm done with code cleanup. The goal is to start animation rendering without exporting OBJ/MTL for each frame, assuming you want to make a 'camera travel' into a static scene. For those who need that, it can save hudge amount of time. I've a reference scene that takes 30 secs to export on my i7 (particles+subsurf), which means a camera-only anim of 30 secs @ 25fps would render 30x30x25 seconds faster, ie 6h15min...
- make people understand/believe Blender is a working solution for professional users and that open-source addicts like me are willing to pay for good commercial software such as Octane
My main concern is to avoid as much as possible code regression as I understand some people need the script to save time in their workflow and at the same time want to have the benefit of 2.56 instead of using the stable couple Blender 2.49+official script.