[EM] proxy system

[James Green-Armytage]

>I've actually spec'd  
>up ideas for forming such an online community, where people just simply
>join up and start selecting each other as proxies. 

	Yes, one of the most logical ways to gain interest in direct democracy is
to create an actual forum for it that is capable of expanding to
large-scale participation. I've thought that this would be a pretty
interesting idea for an NGO. If somebody can get the funding together, I
think that they should do it. I would only hope that they use one of the
better available versions of the proxy idea.
	Ultimately I'd want the voting to change hands from this group to the
government itself, in an official and secure procedure. So the long-range
purpose of the group would be to inspire / pressure the government to do
this.
	I suppose that there are probably a lot of internet polling groups out
there already; I wonder if any of them are in a position to expand in this
way. Perhaps one will eventually emerge as being the de facto public poll.
Hopefully if this happens, they will be open to using a good proxy system
(one that uses Condorcet's method, proxies of proxies, and so on), and
they will be interested in pursuing a larger direct democracy agenda. 
	I completely agree that this is an important area for the election
methods movement to invest itself in.
>
>By the way, James, I feel like I have an abstract sense that there is a  
>flaw in your paradox resolution scheme.  

	I thank you for taking the time to read my proposal. However, I disagree
that such a flaw exists; I think that you have misunderstood.

>I'm not sure why, it has something  
>to do with your solution have a time-based judgment to it, when the
>voter's
>ranked preferences all exist at the same time. 

	No, the web of proxy-preferences is static, and each vote is dealt with
separately.
>
>
>Who is A's proxy in an A->B->C->A loop?  You have two rule candidates.
>#1: it would go to A's second choice (A').  So the "settled" version is  
>that A's proxy is A'.  

	Correct.

>This means that B and C's proxies would be A'.  

	No, here you have misunderstood. To be clear, let me repeat my main rule
for loop resolution: "A vote shouldn't travel the same proxy path twice."
	I take it that your example shows the first choice proxies of the 4
voters A B C and D, and that their second choices are other individuals A'
B' C' and D'. 
	In your example, A's vote would travel the A-B path, the B-C path, and
the C-A path, but then it would be prevented from traveling the same A-B
path again, therefore going instead to A'. However, the paths themselves
remain unchanged. The A-B path is not eliminated, it is only exhausted
with respect to that one vote.
	B's vote would travel the B-C path, the C-A path, and the A-B path...
then it would be prevented from traveling the B-C path again, and would
instead go to B'. Likewise, C's vote would go to C'. I hope that this
clarifies the solution for you.
>
>I don't know much about proxy yet - do other proxy schemes already have  
>the proxy-loop problem solved in a bulletproof  manner?
>
	I'm not aware of anyone else who has suggested having proxies of proxies,
and therefore the possibility of a proxy loop. If such a suggestion was
made prior to the time I suggested it, I'd appreciate if someone would let
me know. Actually the problem of resolving proxy loops isn't really a big
deal compared to serious paradoxes like the Condorcet paradox. It
shouldn't pose any real difficulty to the advent of a proxy DD system.

my best,
James