Hi Anthony, I've experimented with 3 other ones. The first one was one of the most well established ones. I found that ultimately it was too finicky handling surfaces with very slight errors in them. Like if a closed surface was more than 1/2 a degree off from being smooth it would refuse to construct it. It's also geared more for "enterprise" type higher cost software with high licensing fees. The second does not have those problems but instead it basically has the same problem as my current one where it isn't on track to get the approx 50 or so man years of focused effort needed to do filleting and shelling very well.

The third one is a little different, it's made by a single developer. It has been very helpful, it is the reason why MoI v4 has a new "best in class" sub-d to NURBS conversion function. It's going to help implementing some various interesting new functions in MoI in the future but I don't really see it as a replacement for the current kernel it's more for adding on some interesting new functions.

There is another one that seems promising and that I would like to experiment with in the future too.

- Michael