These numbers "feel" very accurate - I ended up adopting a peculiar behaviour using st3 & atom in parallel, using st as a default editor for anything that likel

These numbers "feel" very accurate - I ended up adopting a peculiar behaviour using st3 & atom in parallel, using st as a default editor for anything that likely is big or will probably be accessed outside of "project scope" like csv, conf, sql, and sticking to atom as dev environment. Need to adapt here and there but couldnt bother to decide for good