normalize docs need expansion

[alice-i-cecile]

When examining Matrix::normalize
, I have a few questions:

1. What exactly do you mean by "normalization" here? Does this correspond to setting the length of a vector to 1?
2. What happens with 0-length vectors (or their matrix equivalents)? Does nothing happen? Does it panic? Do we get infinities / NaN values?

[tpdickso]

1. Yes, that should set the length of the vector to 1. However, it will not succeed in all cases; if the magnitude of the initial vector is too small then I believe the result is unspecified.
2. Good call -- I believe in that case you'll get a vector with magnitude 0 (since its length is the square root of an empty sum.) There shouldn't be any non-real values in the result since the dimension of the input vector remains the same, so it will still contain zero actual scalar dimensions.

These are edge cases that would be good to document.