I don’t think we’re complaining about the same thing, and not sure you understand what mine would be exactly. Although, I have probably mentioned a few details about why Rhino fails to fillet in a few cases so far.
Because going from one program to another makes formats consideration non-avoidable.
IGES isn’t able to output solids, for example, but STEP can – during the export/import workflow between programs. As I already mentioned. So should be obvious why I mentioned them, especially cause I haven’t seen anyone bring it up yet really.
IGES does have the option but requires “stitching” and can cause more potential of translation errors than just using STEP/STP.
Technically I agree with you, but at the same time I know enough already about geometric entities in CAD programs to know the term “fillet” is relative like many things are.
So, a ‘fillet’ derived from meshes, or surfaces – are two completely different things.
And within each of those two geometry types, a “fillet” can be comprised infinitely many ways.
However, there are a few intended ways which are much more efficient than others – of course.
A user that blindly expects software to be magical and do things automatically via imagination and lack thereof, is a user that will not be as efficient as one who knows the difference between geometry comprised on one entity type and another.
Not to mention those infinitesimal ways to fillet via Sub-D, etc.
I was kinda quoting Einstein, and yes once a 3D modeler that combines all those modelers together in one, is understood, it would be obsolete. 
And to kinda quote Musk, ‘some technology shouldn’t exist’.
It’s important, imo, to reduce redundancy. There’s many other programs we could mention, but at some point we’d have to be more specific about GUI features, rather that broadly merging programs together.
And as technologies become obsolete, like when patents expire for example, then you’ll see different programs actually embodying the same features – or even sooner via licensing.
I’m glad you mentioned that. I’ve seen it before. It’s a really good read.
I especially like how Mark said, “… Many people are disturbed by the fact that if they trim a solid (like the wall of a building) they reveal the inside of the boundary representation (BRep ) of the solid. This is more to do with the technique used to cut the wall than an issue of “was it actually solid”. If you cut a solid using a Boolean operation, using either a surface or a solid as the cutting tool, you will never see inside the BRep. So everything is kept secret.” 
What does the author mean by “see inside the BRep”. It’s kinda misleading. Maybe the author is referring to the inside of the multi-faced-BRep-solid.
I think part of the problem is “everything is kept secret”.
This ultimately will prevent users from actually knowing how to overcome errors and bugs that occur from time to time, that are associated with geometries comprised of certain “secret” format compositions.
But yes it would be nice if things were to automatically design themselves out of thin air like magic flawlessly and instantly – as a user I could wear a blindfold.
