Normalizing a multivector is currently dealing with weights on degenerate blades as if they were not degenerate. Shouldn't work that way.

[adiffer]

Normalization is handled at the level of the weights and probably shouldn't be because weights for blades that square to zero shouldn't be included in the coefficient sum.

[adiffer]

I have renamed the method to .normalizeOnVS() since it IS a kind of normalization.
I have started on another .normalize() method that makes use of what most people seem to do. They divide by the square root of the product of a multivector with its reverse. Mine will take U -> U/ sqrt(<reverse(U) * U>_0) for now. I have to take a peek at what Shirokov would say on this, though.

[adiffer]

Unit tests for new method are passing now. Still gotta check Shirokov.