Download the .pdf. Australia currently uses 18th century voting - paper ballots filled in with a pencil. That however has given us the best democracy in the world for 200 years! I do not think anyone would suggest that any election ever has been stolen from the people in this country by voting irregularity. By lying politicians yes, solutions for that one are most welcome.
However the cost of doing things this way is getting prohibitive, and the logistics are such that errors are creeping in. It really is not good enough that the result of Government took 2 weeks, and the result of the senate fully a month.
So we do need electronic voting.
The problem with electronic voting is this worrying thing we are seeing in the USA where reports are made of machines coming with the result pre-programmed, or of a machine malfunctioning and spitting out the result before the election. Contracts are being awarded to companies owned by the Billionaires that run the world, what could go wrong with that?
The answer is simple, electronic voting with an audit trail that does not identify the voter. This can be achieved by simply using a voting machine that is currently available as a point of sale device - an A4 tablet with printer attached, and connected to a network hub and then the Electoral Commission computer. Votes can then be announced soon after polls close in the west.
- Each vote is identified with a random 6 character alpha code, providing 308 million combinations;
- Each voter receives a voting slip with that number and details of when and where the vote was placed;
- Voters can log on to the AEC website, key in the number and see how they voted - to make sure their vote was not changed;
- In case of the person being mistaken about how they voted, the vote is encoded in a barcode at the bottom of the slip using an encoding that is available to someone we trust like the Privacy Foundation.
Hash Bar Code
The hash barcode is the extra layer of proof. A voter who sees a different vote on the screen than they recall can take the slip to the AEC and ask them to print the vote off the barcode, then pull the vote out of the computer off the voting reference and compare. If there is an irregularity then we have date, time, machine and intended vote to look into who cheated and also ensure the correct vote is recorded.
Senate and House of Reps
Separate slips would be generated for each as each would be a separate voting action. The size of the Senate poll may require a full sized screen not tablet, these things come in different sizes so running a test vote somewhere will sort out size -vs- useability issues.
Additional Candidate Data
An electronic system would allow for say an information button next to each candidate that brings up a lighbox with a 100 word precis of their policy for proper candidate identification. Photos are possible but worth a conversation - I would think that would encourage cult of personality over policy based voting.
Programmed How to Votes
If you want to vote for a candidate you should be able to just click the link next to their name that prefills your ballot with their preferred voting option. Then you can manually scroll through and change it.
We get to see our preference allocation
Your vote is tagged with a unique code that can stay through the preference allocation process. That means you can log in after the result is declared and see where your vote wound up. Backroom preference deals will be much discussed when people realise their vote for a supposed green party wound up getting a Liberal senator elected. Or whatever, just an example.
There were 3400 polling stations in the 2013 election. We will need around 120,000 machines at a cost of $1000 per machine with supporting infrastructure. Total cost = $125 million. Programming it of course is something else. We know the Obamacare website cost $700 million, however 2 Harvard Students programmed out that same website in their garage in one weekend for free. So there are 2 questions here, how much should it cost and how much will it cost.
Note the 2016 election is expected to cost $250 million, and these machines should last 2 Federal, 2 State and 2 Local elections (5 - 6 years) meaning the cost per election is $20 million. The saving in post election counting should be significant, let alone of course the benefit of accurate and immediate results.
Why 308 million combinations?
Journalists are going to try and pick the result by hitting the AEC website and trying random combinations. This gives them 30 bad tries for 1 good guess (on average). If we limit the number of tries to 10 then the result should remain secure. Some additional security to prevent hackers using rotating ips and number generators would be necessary. Remember though all they are doing is guessing the result not changing it.