There was recently a question about why people lose abilities when paying a bounty. I'd like to ask a related question.
My Super User reputation score is 2,576 (as of this writing). I have answered 159 questions. My average score is 16.2. It can be said that on average I'm going to provide THE answer to any question I answer.
Imagine for a moment that I knew of a user with the same reputation as mine. Only, they had answered 318 questions instead.
I ask, who would you trust more for an answer? My average score of 16.2, or theirs of 8.1?
Now for the really hard question. A users average is 30.4, and their reputation is 1,288. Would you be more interested in their answers or mine?
My point? There is an easy way to make sure that someone that contributes to this site is not penalized for providing a bounty, and the community isn't penalized for them doing it.
Switch scales for acquired functions after a minimum score is achieved. What I mean is this. There is currently a scale that shows how many reputation points you have to have in order to earn the ability to do something. Once that ability is earned, your AVERAGE is what keeps the ability.
An example: I decide to provide a 100 point bounty. My score would drop to 2,476. Now I can't create tag synonyms anymore. This makes me less effective in the community, because I did something to reward someone in the community. My average score is still above 15 for every answer I've given. I'd still be expected to be able to provide THE answer for any question I answered.
By modifying the system slightly the penalty for providing an incentive for answering questions (doing research, causing conversation and interest) is minimized. Yes, reputation is the currency of this board, but it is SILLY to punish someone that incentivizes the community, and even sillier to take away an ability they have demonstrated they can be responsible with.
Saying that, "coming up with 100 more points isn't that big of a deal..." Well, then reputation on this board isn't really worth all that much, is it?
What I'm driving for in summary is that we need to have a quantitative component (we have it, the reputation score) AND a qualitative component. This can be from many things (badges (number, color), averages (mean or median), active time on a board). Just a linear scale doesn't really tell me much (and we all already agree on that, look at the badges we have).