I find your insistance on *MUST* a bit strange. Wikipedia already is a great success by any definition without any artificial rewards system so there's a proof in the pudding that artificial rewards system is not crucial to Wikipedia success therefore is not a *must*.
The problem with your proposition is that it radically alters how the system works. While you're of course convinced that the new system will work so much better, I personally don't see it that way.
First there's a very flawed assumption that just giving people ability to donate will create enough dontations to support the system. Historically we can say that this is not the case. Despite constant fund-rising efforts by Wikipedia foundation, donations barely cover the cost of hardware necessary to keep up with Wikipedia's growth. So I don't see how you can make suddenly turn the trickle of donations into a stream big enough to support not only basic needs of Wikipedia but also leaving enough money for paying contributors meaningful amounts of money.
There there's also this fact that, while the issue isn't completely settled, there is psychological research saying that extrinsic rewards, contrary to naive but popular belief, not only doesn't increase performance, but actually lowers it, see e.g.: http://www.psych.rochester.edu/SDT/cont_reward.html http://www.nwrel.org/request/oct00/motivate.html http://mentalhelp.net/psyhelp/chap4/chap4q.htm
The important conclusion is that there is lack of conclusive evidence that extrinsic rewards actually help and there is psychological research showing the opposite.
Krzysztof Kowalczyk | http://blog.kowalczyk.info