open electoral system (strawman plan)
Michael Allan
I have a strawman design sketch for an 'open electoral system', and I
wonder if it can stand up to scrutiny? In particular, I make some
assumptions about community participation.
The plan is to develop it now, and deploy it as a test-bed in the
community, operating as an advanced poll on official elections.
(Details are still rough, and this was written in haste, please pardon
oversights.)
(I am posting separately to APSA_ITP (social and computer scientists),
and to TorCamp (developers/engineers):
http://lists.hmdc.harvard.edu/lists/apsa_itp_at_lists_hmdc_harvard_edu/2007_08/threads.html
http://groups.google.com/group/torcamp
Open Voter Registry
-------------------
Design-wise, the #1 problem is enforcement of a single vote (to
prevent ballot stuffing, etc.). The solution proposed here is similar
to a traditional, enumerated registry: it's an address-by-address
database of neighbourhood voters. Its difference is, instead of being
validated by a periodic survey, it would be validated by a continuous
web of trust.
Structurally, each entry in the voter registry has:
* land address
* name
* e-mail address
A web of trust then interlinks and validates the entries. Each link
(trust-link) is directed from one entry, corresponding to a registered
voter, to another entry. To become eligible, a voter must aquire a
threshold of trust (say 3 trust-links).
There are also mistrust-links. Each mistrust-link cancels a
trust-link. Voters will use mistrust-links to "point the finger" at
dubius registrants in their neighbourhood, such as houses with too
many occupants, or non-existent addresses.
Functionally, the registry accepts commands directly from citizens.
Citizens send in their intial registration, and maintain their
trust-links, by email (verified by challenge/response; later, a Web
interface too).
The full registry will be shown on the Web, complete with all trust
links, etc. It will be especially easy for a voter to browse through
her own neighbourhood, and adjacent neighbourhoods.
Neighbourhood Registrars
------------------------
These will be unofficial volunteers, living in the community, who are
especially active in assisting their neighbours with the registration
process, and with policing the web of trust.
For registration, the registrar might visit other people in the
neighbourhood to explain the purpose of the open electoral system. If
they give permission, the registrar could then proceed to register
them, and grant them an initial trust-link. (People could
self-register too. They would begin with zero trust-links.)
The registrar could then send each new registrant an email, explaining
how to cross-link with their near neighbours, and thus bring an entire
neighbourhood into eligibility.
The registrar would be especially active in policing the resulting web
of trust, in her own neighbourhood. She would often be the first to
point down abusers of the system. In all of this, she would set an
example for others in her neighbourhood.
Voting System
-------------
Once a voter is registered, she can begin voting for all government
offices in her jurisdiction. End-candidates need not be registered,
but must register in order to cast their own votes (and thus channel
the cascade to higher candidates). All candidates will be chosen by
the voters, simply by the fact of voting for them. (There are no
'nominees', all citizens are eligible candidates.)
Votes (and vote shifts) are sent in directly from citizens by email
(later, Web too). The current state of each elections will be visible
on the Web, complete with the delegate heirarchy. Voters will thus
know where their votes cascade to. (This is a 'delegate cascade'
system, as we discussed earlier in APSA_ITP. People vote 'close to
home', based on their own knowledge, and on information they trust.)
A single voting system is used for all levels of
government. Candidates are identified by email address. (The same
voting system might be used for other purposes too, such as ad hoc
polls. More important, it will later be used to vote for draft bills
in community legislatures.)
Problems
--------
Vote buying: Can a law against that be enforced, or not? If not, we
need to hide all or part of the cascade from the public (so illegal
buyers cannot easily verify the voting behaviour of their purchases).
But that solution would introduce problems of its own (complexity,
policing the hidden parts against fraud, all the problems of secret
ballots in traditional voting).
Some people will not want their addresses posted on the Web (web of
trust or no). However, back when people were listed in the phone book
-- address and all -- most people accepted *that*. And this is a
community intiative (not top down) so people may be more receptive to
it. And, if it makes neighbours more aware of each other, would that
not make the community safer? So, hopefully, not too many will opt
out.
Jurisdictional scale: Is it a single national system? Or multiple
regional systems that cross-talk to calculate the cascade? Cross-talk
complicates things. (But consider, international elections and polls
etc. would necessarily require it, anyway.)
Disenfranchisement of voters lacking computer/network access: Maybe
the community needs to come up with ideas. Since delegation is built
into the system, it should not be too hard to use it, at least as a
temporary stop-gap; e.g. sneaker-net between voter and trusted
delegate in the neighbourhood.
Over-representation of residents ineligible to vote in the official
election, e.g. non-citizens. In some types of community, this might
significantly skew the 'poll'.
Summary
-------
My main concern is on the social side: Am I assuming too much of the
public community here? Especially in the registration process, will
they pull together and participate? (I can't forsee why not. We could
assume political leadership. But I have been cautioned against
assuming too much of a community, especially in new tasks.)
On the technical side, we could develop and deploy this system fairly
quickly:
* email challenge/response (easy with TMDA)
* database (any)
* a Web view of the registy/elections (bulk of the work is in here,
but nothing fancy is needed, at first)
* glue, to hold everything together.
Maybe we could be ready for registration, in a single riding, in 4
months? (I assume it would be a community development effort.
TorCampers, can I assume leadership/organization from TorCamp,
do you think?)
The significance is, the unofficial system could have a decisive
effect on official elections. When fully deployed, it would offer a
continous, around the clock, 'poll' -- showing top candidates, and
likely winner -- well in advance of the next, official election. Come
the election, if the unofficial candidates ran, they would start with
full public backing. If the top candidate held onto it, she would
likely win.
--
Michael Allan