Hello Hacker news, feel free to fight me on this, just understand that you're wrong and I'm right; the number one way to stop burnout on a project is: stop brea

Hello Hacker news, feel free to fight me on this, just understand that you're wrong and I'm right; the number one way to stop burnout on a project is: stop breaking backwards compatibility. For reasons I cannot quantify or qualify, projects that go through rapid transformation and repeatedly break compatibility with themselves burn like comet, bright and intense, before everyone jumps ship to the next hot thing. Projects where people stick around know that real world users depend on them and they cannot change simple things like I dunno... renaming the boot drive from /boot to /bootfs (yep, looking at you Raspbian, and yes, that was fucking stupid). Everyone reading this thread will be outlived by a COBOL program running on ZOs, but their ridiculous node contraption will be forgotten by next year (and it won't build next month anyway). Causation/correlation/selection bias... I think the simplest thing wins, which probably already is done