I've heard this argument before, but unfortunately, it's not true. I'm not saying that empirically, Mac users aren't faster at using the menubar than Windows users-- I don't know, I haven't tested them. But the large differences you speak of would not be accounted for by Fitts' Law. See, Fitts' Law is inherently designed to account for movement in 1-dimension. That doesn't stop people from applying this law to movement in 2 dimensions, even though the law only allows for one "width" parameter. How can they do that? Well, there are a variety of ways, several of which are outlined in MacKenzie & Buxton's influential paper "Extending Fitts' Law..." What they found, comparing several different methods, is that the law is most accurate when using the _smaller_ of the two sides of a rectangle as the width parameter in Fitts law. Thus, while one dimension of the target (menubar) is effectively infinite in width, the other is actually quite small. While there may be differences in pointing time by making the larger side ever larger, these differences are not accounted for in formulations of Fitts' Law, and in general the RT differences are not nearly as dramatic as those described here.