I don't mouse over to the corner of the screen in all but my laziest moments; but sometimes I happen to only have one hand on they keyboard (perhaps I was just closing a tab in Firefox, or upvoting a comment on Hacker News?). It so happens I'm right-handed, and so my mouse is on my right side, leaving my left hand in the perfect position for the windows key.
My common experience is to press the windows key and mouse around to click. While I proudly tout myself as a keyboard-primary user (and this is still the case. I use the keyboard as my sole tool in many more cases than some of my co-workers), I still happen to use the mouse from time to time and this is an example of it.
Also, when you're using an interface that has such accessible buttons, you tend to learn the general area of buttons, so it's not so much "scanning" as it is "oh yeah, that one repository is about here on the screen, let me swing my mouse over there and click it". The large buttons, again, help make that easier;* so, I don't quite have the same experience.* *
For my less-commonly used applications, I do follow the [windows key] enough text to load a program, usually the whole first word in its name because I type around 90-100 wpm and it's easier for me to type words than sub-sequences of letters [enter]; but it's rather surprising how much I've found myself actually adding apps to the start menu because I find clicking easier. Again, this is just from my personal experience, YMMV.
* I'm not a usability engineer, but I've read an article or two that talks about how difficult actions are for a user, and precision vs speed comes up.
edit:
* * As an example of this experience in progress, you probably know where commonly-clicked icons are on your desktop, if you happen to use your desktop to hold frequently edited documents or shortcuts. If someone re-arranged your icons, you'd be slower because they would no longer be in their previous, familiar locations.