Over time I realized a very annoying property of the stackoverflow reputation system: Easy answers to popular question are rewarded with a lot more reputation points than actually great answers. In case anybody actually needs examples for this, here are two examples from my own profile:
This answer gained me 420 rep, I invested 2 minutes writing it, and it only was a wild guess:
This answer took me about an hour for the research and the math and gained me 40 rep.
Now I know that you are bound to get more rep for answers that are useful to more people, and I'd say that’s fine. But the current situation is quite extreme: 30% of my reputation is from 5 of my 330 answers, and they are definitely far from being my best ones.
A final example that is not from my profile, but more extreme: Do you really think this answer is worth 40000 reputation?
My proposition to improve that situation would be to make the reputation gain from a single answer non-linear, maybe even have it capped at a certain point. I’m not sure on how that non-linear function should look exactly, and would leave good propositions for that to users with more experience than me.