Rules of thumb for designing Electronic Direct Democracy systems

[metamerman]

This is an informational post related to the issues raise in the critique:

• Note that these rules relate to technical issues. There are a parallel set of marketing issues (features that will make it easier to market the final system to The People). Unfortunately these two sets are in direct opposition in many cases. For example, "dictatorship by demagogue" is a particularly easy sell to our Pleistocene-tuned brains which is why Hitler was "elected" dictator by 88% of the German population in a 1934 referendum. Unfortunately the outputs of systems that are easy to sell to end up being far worse for The People, an outcome that apparently in most cases they seem to be unable to predict.
• The attraction, and indeed the viability, of democracy is due to the promise that everyone gets an equal vote in every decision that affects them. An effective EDD system must provide the effect of this equal weighting regardless of the actual participation rates: That is, the final decision should be the same as if each individual that will be bound by the decision had unlimited time, knowledge, and motivation to ensure that they cast a "correct vote", the vote that best represents their fully considered preferences. Failure to ensure this outcome means that the resulting system will be an oligarchy, not a democracy (and virtually all existing and previous so-called democracies, including all previous EDD proposals, fail to meet this requirement).
• No matter what you do the majority of The People will not directly participate in most decisionmaking because doing so requires a significant amount of effort (i.e., a time investment of more than a few hours a year). Therefore delegation is not only a required feature of any viable system, it must be the default operating mode (i.e., people must "opt out" of each decision if they don't want to be represented by proxies/delegates).
• The existence of political parties, demagogues, celebrities, "supervoters", or any other type of aggregated political power works directly against the requirements of Update README.mediawiki #1. Therefore any proposed EDD must not only not facilitate this type of aggregation, but must include features to prevent it from arising spontaneously.
• There is almost no scientific evidence that a secret ballot is a necessary feature of a direct democracy other than as a marketing feature. If rules 1-3 are followed even the potential benefits of a secret ballot disappear. Because a secret ballot effectively eliminates transparency and the ability to have a verifiable delegation/proxy system, no viable EDD can include any provision for a secret ballot. The need for a secret ballot is this generations "The Earth is flat" or "Disease is caused by bad air": Beliefs that are nearly universally held, yet bereft of actual factual support.
• The requirements of rule Update README.mediawiki #2 can be met by a wide variety of proxy-matching algorithms. This is where opportunities for creatively in EDD systems abound. The other rules, however, are much more restrictive and so present much less opportunity for alternative solutions: Any system that does not follow them will surely fail (indeed in most cases they can be shown to fail by definition, before even a single line of code is written).

As was the case for the critique, if you'd like to discuss these issues, I suppose we can do it here, but there are two other forums that might be more appropriate for it:
The Metagovernment mailing list:
http://metagovernment.org/mailman/listinfo/start_metagovernment.org
The forums on the Matchism website:
http://www.matchism.org/forum/

[dsernst]

There is almost no scientific evidence that a secret ballot is a necessary feature of a direct democracy other than as a marketing feature. If rules 1-3 are followed even the potential benefits of a secret ballot disappear. Because a secret ballot effectively eliminates transparency and the ability to have a verifiable delegation/proxy system, no viable EDD can include any provision for a secret ballot. The need for a secret ballot is this generations "The Earth is flat" or "Disease is caused by bad air": Beliefs that are nearly universally held, yet bereft of actual factual support.

Curious if you have more data to back this up?

[metamerman]

Here's a couple to get you started:
http://www.matchism.org/refs/Gerber_2012_SecretBallot.pdf
http://www.matchism.org/refs/Gerber_2013_DiscussVoterChoice.pdf

There's also more discussion of this on the Matchism site:
http://www.matchism.org/system/

To summarize, Normies not only don't insist on a secret ballot, they don't even believe that we have one now. But again, the real key to understanding this issue and why it's an "earth is flat" belief analog is the transparency issue: Unless you can go in and verify by yourself that your vote was accounted for in the total, nothing else matters because you've already ceded your "sovereignty" to whoever does the final tally (i.e., all this blockchain stuff is just rearranging the deck chairs on the Titanic).

[SFSandra]

The fact that many do not believe their ballot is not secret is demonstrated to have a suppressing effect on voter turnout. https://datasociety.net/output/privacy-security-and-digital-inequality/.

@metamerman can you explain in terms of public and private keys how a voter is giving away sovereignty to whoever does the final tally?

[metamerman]

@SFSandra, sorry, I can't find anything about turnout (or voting at all) in the link you provide or the paper it refers to. Even if it did, I'd consider it evidence that we need a new transparent public decisionmaking process because the problem is that people don't trust the government, not that they don't trust each other. The fact that these are conceptually two distinct things just shows how screwed up our conception of what government is (i.e., it's only supposed to be a way to coordinate implementation of The Will of The People).

As for the sovereignty issue: Are you going to be running your own server that's keeping track of every vote transaction? If so, how much do you think that would cost in CPU time and network bandwidth? Please assume a vote every week with a few billion participants (what I've calculated is what is required to run a local and global-scale government, and what proxyfor.me is designed to accommodate). For reference you might consider the transaction cost for Bitcoin (several dollars and a several minutes of CPU time for each transaction).

If you're not doing, or even capable of doing, your own verification (and for 99.999% of the people that'll be the case), how could you possibly be convinced that your vote is more likely to be counted in this distributed system than in a single (perhaps mirrored) server system with a public vote where you can just go in and verify your vote with any web browser?