Minimum angle of a quad or triangle

[Peter Blomberg]

I'm working on part 35965 and it has an underside periferal wall of thickness 4 LDU and multiple perpendicular branches. At worst, the 4 LDU side branch is 160 LDU away from the corner, thus making the sharp angle 0.0349 degrees (0.0006 radians). This is larger than the minimum angle of 0.025 degrees specified in the official requirements, but it is really long and narrow triangle!

How accurate/tested is this minimum specification?

What makes something work having the 4 LDU side branch up to 189 LDU away from the corner, but suddenly stops working at 190 LDU from the corner?

Can I as a part author choose to allow T-junctions when the sharp angle is less than 1.43 degrees? This corresponds to a 4 LDU high object 160 LDU away from the corner. Interestingly, that angle is 0.025 radians!

[Philippe Hurbain]

I definitely make t-junction in that case (and many others) especially when it saves on the number of surfaces to render!
I never could understand the ayatollahs of t-junctions: if it causes no visual defect and saves some surfaces, it's fine!

[Travis Cobbs]

I definitely make t-junction in that case (and many others) especially when it saves on the number of surfaces to render!
I never could understand the ayatollahs of t-junctions: if it causes no visual defect and saves some surfaces, it's fine!

I'm pretty sure that all t-junctions will be viewable from an angle that causes a visual defect, but the defect can be minor, and also can be invisible from most angles. This is why they are discouraged but allowed.