I think this is why it helps a lot to build something you actually use. Because then, the barometer for what is good becomes a lot more defined: "Did I solve the problem I had", and then slowly build up from there.
Instead of trying to imagine a thing that someone else might or might not need.
I've been slowly chipping away at a heroku alternative called Canine [1] for the better part of a year now on the side, and for once, I don't feel tons of pressure or self loathing for not working on it quickly enough.
I use it every day now, and whenever I come across something that I wish was a little better (at the moment, understanding how much memory is used by the cluster is a pet peeve), I ruminate on it for a few days before hopping in and making some changes. No more, no less. It helps me get away from "what is the perfect solution", to "can i fix this thing that annoys me right now"
replies(2):