Magnus Manske wrote:
But if consensus (why does that word remind me of the Borg? ;-) can not be achieved after years of discussion, voting might be a way to actually get something done...
With the risk of getting overly philosophical...
Even in a parliamentary democracy, there has to be consensus about using voting as the method and how to apply it. Otherwise, the method will be useless. In open organizations (where anybody can become a member at no cost) such as the IETF and Wikipedia, voting is problematic, since it is unclear who is eligible to vote. This leaves a suspicion that the opposite party cheats by subscribing new voters, undermining the consensus about voting as a method.
Within the IETF and in some other places, consensus is often hailed as a better alternative to voting, but they are two methods that have always been intertwined. When the IETF hails "rough consensus and running code" as their successful method, I think "running code" is the more important part, since the obvious problems with voting leave "rough consensus" as the only realistic alternative.