I want to correct some misinformation in TidBITS-139 about Conflict Catcher. The article claims that Conflict Catcher automates the process of loading startup documents one-by-one to identify conflicts. Conflict Catcher does not load INITs one-by-one, but instead loads half of the startup extensions at a time to minimize the number of restarts it takes to locate an INIT conflict. Here's how Conflict Catcher begins to tackle a problem. After you start Conflict Catcher, it loads the extensions that it wants while restarting the computer. When the Finder comes up, you check to see if the problem exists (and that includes opening an application if that's where the problem shows up), and restart the computer. Conflict Catcher will then ask you if the problem exists or has gone away and enables or disables extensions as needed. It usually takes about four or five restarts to pin down a conflict, although that of course varies with the number of INITs you use.
Conflict Catcher is always the first INIT to load, so if the conflict is between two INITs and causes a crash during startup, you can inform Conflict Catcher that the problem exists when you restart the machine. If the problem is more subtle (i.e. the Finder has a corrupt display), then you tell Conflict Catcher when you restart.
Also, Conflict Catcher is able to isolate conflicts between multiple INITs. The article also mentioned incorrectly that Conflict Catcher somehow traces code after startup. Actually, Conflict Catcher only patches a few traps to perform the startup file reordering and to do the ICON wrapping. Both of these features can be disabled so that Conflict Catcher is guaranteed not cause any problems.
[Thanks for the explanation, Jeff. It sounds as though Conflict Catcher will help the user identify and solve conflicts, which is even better than it doing it automatically because then the user will learn from the process as well. -Adam]