the technical side of trust free voting can easily be solved with namecoins (even directing votes to other voters) or colored bitcoins.
The part of the trust-free mechanism is derived from the fact that votes come from the addresses that paid the membership fee with a currency spendable on a global market.
This removes the necessity of maintaining a separate database connecting "who paid the fee" with "who has voting tokens". I need to look more into Namecoins though. Last time I read about them there was a discussion whether they need to be destroyed or not when they are spent, that was probably a year ago.
Could you please describe the voting mechanism as you see it with Namecoins?
One problem is identifying voters while still letting them be anonymous. Even that can be solved by shuffling vote tokens with random people or people you trust.
Regarding identifying members vs anonymity of votes there is indeed at least one problem.
If the system provides benefits to members (pensions, sick leaves, etc) there needs to be a way to
identify who is eligible to receive them. At this point I can think of the following approach:
Split the membership fee into (a) public part and (b) voting part.
Those members who paid the public part of the fee become eligible to the benefits of the system. The voting part of the fee is anonymous and voting itself doesn't make any member eligible to any benefits just influence the voting outcome. I think it would be safe to assume that number of voting members should be less than or equal to the number of publicly registered members. On a local scale community needs to be cautious to an outside influence in the voting process, however on a large or global scale rigging the vote outcome becomes prohibitively difficult cost-wise since underlying monetary system has limited supply.
More importantly the system also needs to have ways to determine if an upcoming question is important or not. It is impossible for all people to vote on every question somebody can come up with (why does Atlas come to my mind?
).
Depending on how many people think a question is important and how clear the outcome looks the system would have to determine the number of people necessary to vote to come to a result. Of course everybody should be allowed to vote if he wants to or does not like the preliminary result.
I agree this would be an interesting challenge. People need time to do actual work too, not just chatting the whole day about how they are going to vote. The basic concept I can think of is that members send proposals of the change they are willing to make to an already existing consensus with a schedule to vote at least one week (or even month) ahead. If within this period enough members approve that the community indeed needs to look at this issue then the proposal is added into the voting queue.
Sort of how WEB2.0 content management and rating system is working today.
PS: on the side note...
Another area where this system can also be applicable is a company/corporation. Right now corporations are modeled after the idea of central governance - there is one chap at the top who issues commands to his "generals" and those in turn are in charge of small armies of developers which normally have no say at all. For people with libertarian mindset it might be hard to accept any managerial authority on top of them who would tell them what to do on a day-to-day basis. The system with provable trust-free way to achieve consensus might be easier to swallow. As an example, imagine working on something like building a Boeing 747. That would require a lot of people from many different areas of expertise to do a lot of work in a very coordinated fashion or it simply won't fly.
