OK, this isn't a normal blog post.
It's the output from a quick and dirty app I wrote, which parses the election result data from the recent Pirate Party Australia National Congress Election. There wasn't a key (that I could be bothered figuring out) to understand the verbose, 2d output from the official results, so I've made my own, extremely verbose, 1d output. It's long, so if you're on the homepage of this site, don't click "read more" unless you really want to. (that ought to drive traffic).
Anyway, the main thing I noticed was that there are some completely missing pieces of votes for motions (though not candidates). I hope next year there'll be an easier to follow audit trail, as that kind of "small discrepancy", while being understandable for a small (but growing) organisation like Pirate Party Australia, would be totally unacceptable to be used in governmental elections.
The other thing that I noticed, which is likely either a usability issue or people getting bored of such a long form, is the large number of people who filled out the motions but then didn't bother to vote for anyone.
I think the election system has been quite a success, and if it can be improved by having voter-verification support, an audit system that can trace down discrepancies without compromising the secrecy of the ballot, and be handed off to an impartial administrator, then we'll have something to use as a positive example when the inevitable debate around electronic voting comes to Australia.
Finally, a disclaimer: Quick and dirty, as mentioned above, means I have not had anyone verify my program. Ask me for the (begging for refactoring) source code if you're interested, or prove the data wrong with your own program if you'd like. I'm fine with being told if my output is wrong. To see the kind of sceptical nature I have, read about my home-conducted double-blind wine-tasting experiment, in which I identified about 20 procedural issues.
And now, the deluge:
