Plan experiments with reproducible study layouts.
Recent revisions are kept. Restoring one creates a new revision.