What if scattering or absorption factors are greater than one?

[fiedl]

Currently, the --scattering-factor and the --absorption-factor are assumed to be between zero and one. If zero, the photons interact instantly when entering the hole ice. If one, the photons interact the same way as in the surrounding ice.

But actually, the factor can be greater than one, meaning that the interaction probability within the hole ice would be less than in the surrounding ice.

• Add these cases to the unit tests
• Make sure that these cases are handled by the hole ice simulation code

[fiedl]

Some thoughts on the geometric cases: When the interaction factors are allowed to be greater than 1 then a switch for the special cases "2b" and "3b" is needed.

[fiedl]

Currently, I'm using nan to indicate not only that an intersection point does not exist, but also that an intersection point is not between the interaction points A and B of the photon.

In "case 3b", I need access to YC, which is not available in this case, because it requires termination_point_ratio, which is set to nan because the photon ends inside the hole ice.

Probably it would be better to not cut off the ratios if they are <0 or >1. This would make all necessary values available to "case 3b" as well.

For the calculation of the number of medium changes, one would need to check whether the entry_point_ratio and the termination_point_ratio are between 0 and 1 rather than checking for nan.

• Try checking for 0 < a < 1 rather than checking for ! my_is_nan(a).
• Tests are still passing
• Does this make a difference elsewhere? -> Yes, the hole ice algorithm becomes simpler in several places (fiedl/clsim@5eb66c6)
• Remove un-needed code from intersection.c