I've done it both ways, and to be honest other than the ability to give the code and webform to two different people (which doesn't happen where I work), I see very little advantage of one method over the other. I can understand how it is nice to have your code in one file, and markup in the other, but I can also understand the ease of having both in a single file.
I don't see how using code-behind enforces any sort of "good" programming, because you can still write crappy code in code-behind, and I've seen some really clean stuff with in-line.
So my real questions, what method are you using? and is there anything you can do programatically with code-behind that simply cannot be done with in-line code? or does it all really come down to the simple matter of "personal preference". Microsoft seems to not take a real stand on one way or the other and performance wise, it seems a wash.
Opinions?