Verifiable and Accessible Electronic Voting System

At Linux Conference 2013 today at the Australian National University in Canberra,  Craig Burton, E-Voting Manager, Victorian Electoral Commission, talked on "Designing a Verifiable and Accessible Electronic Voting System". I had considerable difficulty understanding the presentation, perhaps due to my limited knowledge of e-voting systems. Dr. Burton began by detailing at length the problems inherent with e-voting systems and asserting they should not be used. He explained how paper voting systems are vulnerable to "retail" voting fraud (such as ballot box "stuffing") but not wholesale tampering with the tallied results, which e-voting systems are subject to.

At this point I was more than a little confused as to why an E-Voting Manager would be arguing that e-voting should not be used. Why the Federal and ACT Electoral Commissions would have been using e-voting for the last few elections if such systems where known to be so insecure? Also if computerized back-end systems are very vulnerable, then the current federal and ACT voting systems are at a high risk, as the paper ballots are tallied by computer.

Dr. Burton argued for a system where the vote could be checked through a combination of paper slips and electronic safeguards. However, in my view there is no way to use technology to completely guarantee a secure vote. Ultimately it is necessary to rely on the integrity of the electoral staff.

My confusion increased when Dr. Burton showed a video demonstration of what was apparently a prototype Victorian e-Voting machine. This device was using a clumsy looking combination of a tablet computer and a bar-code scanner.  What made this demonstration even more confusing was a demonstration of its touch screen used for the blind. An interface with a mechanical keyboard (as used for the ACT and Federal systems), which provides direct tactile and sound feedback, would be the obvious choice.

It should be kept in mind that there are many circumstances where citizens are involved in making collective decisions and voting in state and federal elections are not necessarily the most important. As an example, I live in an apartment complex, with its own federated system of governance. An elected committee runs the body corporate of the building and elects a representative to the non-profit company which runs the complex of buildings. These non-government bodies provide the garbage collection, streets, street lights, park and gardens, as well as some of the energy supply and telecommunications. Similarly I am a member of various non-government bodies, and own shares in companies, which provide services to the public. Such bodies hold elections and have more effect day-to-day, on the lives of citizens, than do governments. The voting systems used by these bodes therefore need close attention.