Basically we've added back in explained variance and explained covariance by adding 'loadings' which are normalized weights such that w'w=1 or ||w||=1. Loadings are thus equivalent to their PCA counterparts. We also relabelled factor loadings as 'canonical loadings' in line with online dicussion/literature.
2.2.1
2.2.0
Went down a bad route before. New way is much simpler for AO methods, much more elegant, much more extendable by users.