> This is not a reasonable estimate of the work needed. One of the things that did the most to change my mind on this subject was actually estimating that work

> This is not a reasonable estimate of the work needed. One of the things that did the most to change my mind on this subject was actually estimating that work - specifically, estimating what it'd take in a live and fairly complex app, which is where most teams would have to start. (You haven't lived till you've tried to prototype a complicated custom time series chart control with several interactions, that can transparently render itself across a wide range of sizes and be legible and usable in all of them. You really haven't lived till you've tried to test it!) Another was seeing for myself in Amplitude both the absolute fraction and the trend for how few users would even be in a position to notice the result. > Also I was thinking more of "don't waste tons of memory, making it unable to run on devices with less", which could just be a mandatory rule for everyone's benefit. I agree this is worth more than it's valued. Would such a singleminded focus be to everyone's benefit, though? Most apps might not spend much effort on clean UI restore after suspend, but actual data loss is very rare in my experience, and teams have only the capacity that they do. That makes it a tradeoff between features and other bugfixes, and this; the next request I see in Canny for showing the launch screen less often on app switch will be the first, an absence which likewise leads me to suspect that the wants of users in general aren't accurately captured in the perspective of those who are heavily invested on this point.