Hi PaQ, well one thing that makes Fillet difficult is that it uses quite a few difficult calculations inside of it, such as creating surface offsets, and doing multiple kinds of surface/surface intersections.

Then the part that gets really difficult on top of that is figuring out how to make the corner pieces where different fillets are running into one another. That is especially one of the areas that tends to be handled by lots of special case code.


> I'll love to have, one day, my hand on something like
> solidwork, but I'm not sure I'll like the workflow (my
> brain only accept simple concept).

You would probably not find it difficult to pick up now that you are familiar with how NURBS modeling works. But the workflow in those programs can tend to be sort of more regimented or formalized.

They've become somewhat less restrictive over time, but for example those kinds of "parametric solid modeling" programs used to be organized so you built just one solid at a time in a "part modeling module", and then used a separate assembly module to gather things together, instead of just creating whatever you wanted on the fly in the same environment.

They often times would not allow creating curves that looped around in 3D all over the place, instead before you draw you would go through a kind of setup stage for defining a sketch plane... Stuff like that.

But of course all that was not necessarily arbitrary stuff, it fit well with many kinds of engineering workflows.

- Michael