According to Raymond Chen, this usually happens when software relies on undocumented APIs or undefined behavior. When Windows removes undocumented APIs or changes behavior not covered by the documentation, such apps break and people blame Microsoft. For example, I remember one app used to programmatically find some kind of "Print" menu in Windows' desktop environment and call it to implement printing functionality. When Microsoft changed the UI of the desktop (shuffled the menu items around) the app broke and people immediately blamed Microsoft, even though the app is to blame because the exact location of the menu item was never supposed to be a public contract. If an app is a widely used one, Microsoft usually used to create "shims" which detected funky behavior like that and corrected it to make the app work again. I think they don't do it anymore after Windows 8, that's why apps started breaking more frequently.