When I hear that functionality stops working, my first through is..well what update caused a regression.
If mysterious functionality regression is the most frustrating part of your experience using an operating system...any operating system.. and it sends you back into the arms of another it makes me wonder... as a user are you applying the same update policy fairly across your computing experiences, with the same expectations? If an Apple update caused your printer to stop working or wireless network to stop working, would you ditch it for something else?
Apple osX's updates aren't bullet-proof either. They have causes problems for people with wireless and printing..and even mail applications.
Updates are hard...regressions are going to happen for every single operating system vendor....even Apple...who works its a much narrower landscape of hardware and has to deal with far far fewer possible hardware configurations which limites their exposure to potential hardware regressions they can't test for.
So can we really tell if one vendor is doing a better job than another at limiting the exposure and impact of update regressions? Some people in Fedora-land spend a lot of time thinking about how to do a better job here than we are doing. It would be valuable i think to figure out how we can really start trending how we are doing to form a baseline from which to chart progress. One way to do that is to try to get a handle of our performance relative to other os installs, but to do that we need people who are applying a local update policy consistently...and enough of such people to represent a good mix of hardware configurations. Vendor update strategies aside, if as a user of multi operating systems are not being conscientious about using the same policy across operating systems, you aren't going to be able to make reasonable cross vendor comparisons concerning regression impact. If you hold off on applying updates on your Mac, then you need to hold off on applying updates on your linux installs too.