@Maenad YES.
1. yeah, the statement about the units is weird
2. interpretation of coefficients with interaction terms - Hmm not sure if normalizing/sum-to-zero constraints fixes the issue, but I'd have to think through it and do the math...