Rod, do you recall if we intentionally for some reason had the code resolving dependencies before actually testing the activation condition for an attribute definition? It was a bit non-intuitive but I know there are some complications in that code. -- Scott