Comment by CrimpCity

Dang I was hoping the article would have been about non-traditional higher level fractional factorial designs. Like for example you have an experiment with 3 factors and 5, 4, 3 levels each but can only run 20 experiments.