...why imagine? That's life as a Windows/Linux/Android dev. (Apple is sort of a stand-out because it has vastly fewer installable parts and less versions in the wild such that it's actually possible to test every patch level of every supported version of macOS or iOS at any given time).

But none of that makes push updates right or wrong. The reality is that it's less of a push than it is a pull anyway - in this case the client is asking for updates on an interval, and the server says "yep, there's one for you." The client grabs it and installs it. And it's turned on by default because, for the most part, that's the right thing to do for your users: you'd rather them be on the newest patch level. Hell for IT admins though, which is why it's almost always a feature they can disable at will.

So here's where this case differs: it's an "experiment" that's actually just marketing trash pushed through the "experiments" channel which is also armed by default, rather than a security or product update (which anybody reasonable can argue should be on by default - secure by default is the goal, after all). The only "experiment" in this case is seeing how many users will put up with Mozilla continuing to pimp out Firefox to the highest bidder as a grab for a new revenue stream before they reluctantly switch back to Chrome.

And judging by the backlash on patches like this one, it's not going so well...

I'm asking because debian and backports are doig exactly that: separating security patches from the rest, not for a browser but for a whole OS and every applications including firefox.

also this xkcd is not relevant. the point here is that mozilla has quite a history of breaking userspace earning them the reputation of "making far-reaching and very short-sighted decisions in a vacuum."[1]

[1]: http://forums-test.mozillazine.org/viewtopic.php?p=14736466#...