for every week a questions remains unanswered, give it 10 bounty.
This means that questions which have remained unanswered for a year could give you 510 reputation (given that it is also accepted of course)
This system works like an auction, where the pressure to answer a question gets higher and higher until eventually someone will answer it, no matter how difficult the question was.
There is the potential for some really wicked unintended consequences here.
While most questions that are unclear, deeply opaque, or based on horrible misconception get caught and either closed or have the issue explained to the OP so that they get corrected, it is likely that a few slip through the cracks. These questions are not contributing to the community and should not accumulate rep.
The system proposed can not distinguish those from really hard problems (which, perhaps, do deserve some recognition).
Put me down as against it.
+1 I like the reasoning behind the idea. If SO cares about the vast number of unanswered questions, something definitely needs to be done in that department. I think there may be too many practical problems though.
Many unanswered questions are abandoned, i.e. the user who asked them is not likely to return to SO. You would have to introduce community question closing at the same time for this to have any effect - I'm not saying I'm totally against that, but it would be a major change to the system with a plethora of issues.
Also, a bounty would put considerable pressure to close the question. A time limit to decide who gets the bounty would be necessary to avoid disappointment among people putting energy into answering the question. On the other hand, there may still not be a right answer, simply because the question is so hard, nobody understands it, or it's badly worded. What do you do then? In a normal bounty, there's the 7 day time frame after which you have to make a decision. What kind of time frame can you enforce on a year-old question? What if the user logs on to SO once a year? And so on an on. It's going to be too difficult to implement, I think.
