Electronic voting (e-voting) offers some advantages for a complicated electoral system like Ireland's PR STV system. However, unless specific safeguards are put in place, then the complete lack of transparency associated with e-voting poses very serious risks for the integrity and public confidence in the election system. The problem and some possible solutions are discussed in more detail in my submission to the CEV.
One way to add transparency to e-voting is with a system known as Voter Verified Paper Ballots (VVPB). The principle behind VVPB is very simple. The electronic voting machines produce two outputs. First, they store the vote electronically as before. In this form, the vote can be counted easily by computer but it cannot be verified by the voter (because it is electronic, it cannot be seen). Second, the machines print out a paper copy of the ballot as part of the voting process. The voter verifies the paper ballot before confirming (casting) their vote. The paper copy of the vote cannot be counted (easily) by computer but it can be verified by the voter. An audit of a random sample of votes is then done routinely as part of every election to verify that the electronic votes as recorded by the machines is the same as the paper copies as verified by the voters. This audit effectively certifies the officially published electronic vote file, which is used in the official counting system. Of course, the counting system has to be verified as well. This is simply a matter of interested parties running the count on the official vote file using their own independently developed counting software. If multiple counting systems produce the same results from the same input then there is not likely to be a problem with the official system. Here is a more detailed description of how the Nedap/Powervote system for Irish elections could be enhanced to support VVPB (Voter Verified Paper Ballots).
One criticism of VVPB systems is that they negate some of the expected benefits of electronic voting. The original point behind e-voting was to eliminate paper from the voting process and it appears that VVPB puts it all back in again. This is clearly a flawed argument unless the integrity and trustworthiness of election systems matter less than their administrative efficiency. Nevertheless, it does indicate that improvements to VVPB are desirable.
One promising approach has been the development of receipt based systems, which allow voters to verify their votes (without harming ballot secrecy), and which ensure the integrity of the election but without the administrative overhead of managing and handling paper ballots. These systems are based on complex cryptographic protocols, which are at a relatively early stage of development, and as such, it will be some time, before they are ready to be deployed.
My submission to the CEV describes one such system devised by Dr. David Chaum. A more detailed description of it can be found in this article. Another system is being developed by VoteHere. These systems have been criticised as being too complex, and beyond the understanding of the ordinary voter. However, I would argue that this problem is overstated, and is at least partly as a result of their novelty. The fact is that not everyone needs to understand technology in its minute details. So long as a system can be represented and understood at different levels, and that anyone who wants to find out more is not prevented from doing so , and of course that the system is open, transparent, externally verifiable and built on sound scientific principles, then there is no reason not to trust it. Many people ask why bother with this complexity, when the existing paper election system is quite adequate. Of course, this is a valid argument, and any e-voting system must have a defined benefit, that is worth the cost.
Some of the possible benefits of future e-voting systems could be:
Bunreacht na hÉireann (The Irish Constitution) specifies our election system as "proportional representation by means of the single transferable vote" . The exact details are regulated by law, but the system for Dáil Éireann (House of Representatives) is more or less standard STV as described by this article. The following wikipedia article gives more detailed background about STV.
The Dáil system is true STV in the sense that the votes are single units, and when transferred as part of a surplus they are not split. Each vote either remains with the original candidate or is transferred in its entirety and now belongs exclusively to the receiving candidate.
The system for the Senate is also STV but uses the Gregory method, whereby ballots are distributed fractionally, ie. votes can be split among multiple candidates rather than allocated randomly in units. This system is used in the Senate because the electorates are extremely small and more vulnerable to the statistical error caused by random surplus distribution.
A report of a typical STV election count (the 2007 Dail election count for Dublin North) is here.
There are a number of free STV counting implementations available such as this one.
Another system being developed by me is specifically tuned for the Irish counting rules and which uses the CSV file format published by the returning officers for the pilot electronic elections held in 2002. You can download this software, and try it out.