This is an interesting post. The main advantage of multitasking in the OS is to simplify the developer's effort in improving the user's experience. It really boils down to remembering where you were and handling changes that might happen when the user is off doing something else. With multitasking, you assume that you're still alive, so you really don't have to do anything special as a programmer.
In the Palm OS model, the programmer has to save the state when a user goes to another application (remembering where you were). If there is any activity that can take place behind the scenes (for example, an IM app that wants to stay connected when the user goes to the calendar), then the developer has to have some mechanism to handle incoming data from the server (and possible send it an occasional message saying "I'm still here"). Lots more work for developers.
You don't know how often I've had people tell me you can't keep a communications port open while you are checking the calendar on a Palm OS device and then I whip out an app that is doing just that! They then start to tell me that I must've gone through hell to get it to work. Yeah, it was hard the first time, but now that I have the code it's not so hard!
