When not much resources available one have to be more careful (about not accidentally creating O(N^2) where O(N) would work and not wasting RAM on multiples cop

When not much resources available one have to be more careful (about not accidentally creating O(N^2) where O(N) would work and not wasting RAM on multiples copies of the same data) and at same time one have to refuse adding non essential features with bad HW resources / user value ratio. In corporate software it is often easy to find some low hanging fruits allowing to reduce CPU/RAM usage by 10x (but incentives stacked against doing this). In open source software like KDE/Gnome I think it would be hard to find blatantly inefficient code and the main reason is the overall complexity which is hard to reduce without sacrificing some features.