Factor analytic covariance structure updates during REML about sommer HOT 1 OPEN

Yes, that was the original intention but I couldn't find in the literature how exactly they do that. Do they produce fitted values at every iteration and calculate new loadings using the fitted values? or do they use the BLUPs? Most sommer machinery is based on the papers from Van der Werf, Karin Meyer, or Jensen. The papers from the ASREML team are too difficult to understand, like leaving gaps intentionally for people to don't recreate their machinery. Long story short, not actively working on the package now but yes, I plan to put that approach at some point. Any clues or suggestion of a paper that does a decent job explaining methods is always welcome. Also if you can code something please feel free to make a pull request and I will add it :)

from sommer.