Are you ignoring my comment because this problem is actually not a bug but instead a deliberate elimination of functionality?
Let's summarize some facts:
- The tessellation algorithm in Alias works just fine. It is used by the Diagnostic Shading and the Hardware Shade and both work as expected. (They both have the same tessellation settings as the FBX exporter so it should be the same code).
- Tessellation in OBJ Export works perfectly. It has slightly different settings so it might be a different subroutine.
- Tessellation in the JT exporter also works fine but it also has slightly different settings.
- Tessellation for the STL output is also okay but again a completely different interface.
(Why do you have so many different tessellation variants in your code for the exact same task? That's not very efficient...)
- The FBX exporter itself also works except it doesn't tessellate. But NURBS and curves are all exported perfectly.
So, as it seems, the only thing that is broken, is that tiny little checkbox that tells the FBX exporter to tessellate the surfaces.
Normally checkboxes don't break down between releases as they are just an IF statement calling or not calling something else. Even if they would, they could be fixed in a minute. (As the actual tessellation code in Alias is still working flawlessly.)
I'm not sure if you are using an internal translator routine or calling the external FBX translator (AlToFBX.exe) from inside Alias but I guess you use the external one as it has the exact same problem.
(If not, then what is the chance that both the internal as well as the external translators are broken "accidentally" the same way?)
Why is FBX so important in Alias?
Because currently it is the only convenient way to export polygon data out of Alias to an efficient, binary file with all characteristics (shaders, layers, meshes, animation, etc.).
None of the remaining file formats are able to export all of that data at once and tessellate/organise the resulting polygons at the same time.
So, if that tiny little glitch is not fixed within a day and mentioning it after a reasonable amount of time (3 months!) only gets ignored, then the only reason remains is that it's not a glitch but a deliberate castration of the software to disrupt emerging workflows which shift certain tasks to better performing and cheaper software. Such a behaviour would be by the way a deliberate fraud on customers.
Correct me if I'm wrong, please. (By simply fixing the problem.)
PS: repairing and distributing the external translator alone would already solve the problem and would be an easy and quick fix. (No need for a complete redistribution of the complete software.)
So, what is stopping you to do so?