I can’t really check at the moment because it appears to be buggy, but the transformpoint message when used with fluid.normalize~ and fluid.standardize~ appears to be destructive?
As in, you only specify the buffer you want to transform, and it (presumably), does it “in place”.
Don’t get me wrong, I really prefer this kind of approach as it means not needing to have individual buffer~s for every step of the process, but interface-wise, it differs from how fluid.pca~ handles the transformpoint message.
For the dimensionality reduction stuff, the output will (more-than-likely) be of a different size than the input, hence writing to a new place makes sense, but just makes the usage a bit confusing since you need a second buffer~ for some objects and not others.