@C.E. I don't know but in jstor.org/stable/2325085 (1992), equation (1.2), Knuth proposes (1+z)^n = \sum_k {n \choose k} z^k , where the sum ranges over all integers k, which gives 0 for n = -1, k = -1. Others have extended the binomial via the gamma function identity, which gives 1 for all n=k. It's a combinatorial vs. analytic argument. As I said, I don't know that's what Knuth has in mind.

@LukasLang XD Thanks

@MichaelE2 I grinded with nested grid and unmanageable Dividers settings for this one but I will take a look at that if I have to do this again, thanks.

Bug (copying and pasting the graphics may cause a FRONT-END CRASH), three variations:

DynamicModule[{a = {0, 0, E}},
 Graphics3D[{Cuboid[]}, ViewPoint -> Dynamic[a]]]

DynamicModule[{a = {0, 0, 3 Sin[1]}},
 Graphics3D[{Cuboid[]}, ViewPoint -> Dynamic[a]]]

DynamicModule[{a = {0, 0, Sqrt[13]}},
 Graphics3D[{Cuboid[]}, ViewPoint -> Dynamic[a]]]

It seems exact numeric viewpoints are disliked by Dynamiic

DynamicModule[{a = {0, 0, E}}, Graphics3D[{Cuboid[]}, ViewPoint -> a]]

DynamicModule[{a = {0, 0, N@E}},
 Graphics3D[{Cuboid[]}, ViewPoint -> Dynamic[a]]]

I should have mentioned the first three show pink highlighting and have an error of the form

Viewpoint {0, 0, E} is not a triple of numbers or a recognized symbolic form.