Hi Burr,

re:
> Is there any "robustness" left in the geometry kernel regarding Offset\Shelling that you just haven't
> implemented? Or are we at the limits of what's available...

Not really - it's got some fundamentally broken assumptions in it in this area. I think it could be a possibility to write a new one from scratch (unlike fillets) but it might be better to try to get this area better by integrating a different geometry kernel which I have to do anyway to improve fillets.

- Michael