>I bet smaller memory use and snappiness correlate You bet, but do you have any experience of any kind to back this up or is it just a vague hunch? In my anec

>I bet smaller memory use and snappiness correlate You bet, but do you have any experience of any kind to back this up or is it just a vague hunch? In my anecdotal experience, the snappiness of a DE is given almost exclusively by the compositor and amount of CPU cycles and disk I/O ops used by the DE, and less about the RAM used. Of major importance is the input lag of the compositor, with Wayland feeling far more snappier than X11 in most cases. So I wish more of these tests would investigate CPU usage and disk I/O, rather than these RAM tests where users come to the premature "oh look, DE_1 at idle uses 200MB less RAM than DE_2, so it must be faster" conclusion, which couldn't be more false in the real world usage scenarios of how human users perceive snappiness. There's much more to this than just idle RAM usage. So to contradict you, no, idle RAM usage does not in any way always directly correlate to perceived DE snappiness.