Non-modgit model interactions such as tests, critiquing, and scenario management were largely effective, but very ad hoc. Authors tended to have fuzzy goals in mind when they approached tests and critiques, and most failed to plan for multiple scenarios.