Oh okay. I did know about that weakness of PCA. But it seems like in this case letting one prime factor correspond to one unit of distance is a nonarbitrary sca

Oh okay. I did know about that weakness of PCA. But it seems like in this case letting one prime factor correspond to one unit of distance is a nonarbitrary scaling, so I would expect PCA to give sensible results.