This is my attempt at programming a solution to the well-known zenix banana orientation problem (positio musa acuminata problema zeniximus), which suggests that positional variation in the set of all bananas (including, but not limited to both right- and left-handed) is sufficiently complex such that no single banana cutter shall be adequate to cut them.
Unfortunately, my attempt (while still serving as an improvement on the original) was unsuccessful, in that introduced a new problem of positional accuracy, leading me to postulate that the absolute position of a single banana in an isolated physical system is infinitely and unprovably complex, regardless of the triviality of the axiomatic system within which it is defined.
Oh, and I added a spork.
NOTE: I believe this version solves the problem of upside-down bananas, provided the correct rotation is chosen.
1) Measure the exact orientation and handedness of your banana. (Unfortunately I can't provide any guidance here, because this is where I got stuck.)
2) Enter values for "rotation" and either true or false for "leftHanded." The default value fits the position of the banana that I ate this morning, but that's gone now so I can't include it as an example.
If you try the OpenSCAD file, be sure you include the banana_cutter.stl file (it's the original STL, just renamed) in the same directory.