Posts by Matthew Poole
Last ←Newer Page 1 2 3 4 5 Older→ First
-
Using machines to print out ballot papers, which are then OCR'd during the initial count, is fine. Using machines to tally votes right from the get-go is just begging for all kinds of fuckery and skulduggery. The former would greatly speed up counting, since you could be pretty certain that ballots would be in the correct form, and I think it's something that could be investigated here.
-
And on an entirely-related note: WaPo is reporting that Maryland and Virginia will be going back to the old ways after Tuesday's election.
"The battle for the hearts and minds of voters on whether electronic systems are good or bad has been lost," Brace said. The academics and computer scientists who said they were unreliable "have won that battle."
A victory for accuracy and verifiability over the demands for instant results.
-
but I'm not optimistic that they'll happen based on Obama's comments regarding ... the likely protectionist reaction in the US to the global recession.
Obama's a protectionist. That's not news. He's been one for the duration of his campaign, credit crunch or no. And beating some decency into the US domestic lobby groups is one of those "generational timeframe" things you mentioned. Hypocrisy and naked self-interest are the name of the game for Americans when it comes to global trade, and Obama's quite happy to keep that up. It's about the only aspect of McCain's policies with which I agree unreservedly.
-
Again, paper works quite well for this whole voting thing. The #1 requirement for a voting system is surely 'getting it right'.
Unless you're the US, in which case the #1 requirement is "Keeping t3h meedja happy" by giving them a result as quickly as possible. Accuracy is entirely optional. </cynicism>
-
Kyle, if it was just a mouse-based interface, you'd be right. But touch screens add a whole other layer of potential for fucked-uped-ness, because they have their own software for interpreting where on the screen the user has pressed. Mice don't decide how far you've moved and where you've clicked, they just do their thing and let the far-end software make the decisions.
-
Has anyone else noticed the political compass for the US election - the 2nd graph shows a good representation of a map of New Zealand.
I made much the same observation months ago, when it was still just candidates for the Primaries. It's an uncanny likeness.
-
If computer game designers can tell when my avatar in a shooting game hits the bad guy, when both my avatar and the bad guy are moving around, I'm not sure why they can't tell when a person is pushing a static button.
Because it's not a real button. It's a series of coordinates on the presented screen, that must coincide with the coordinates the screen returns. If the screen is miscalibrated, it will return wrong coordinates. If the returned coordinates don't happen to intersect at all with the "button" you thought you were touching, the result won't match your expectations.
The software remains correct at all times. What becomes incorrect is the touch detection systems of the screens. If they become sufficiently incorrect, no part of the displayed button will intersect with the coordinates the screen is returning. -
In a functional democracy, all these matters are academic. But it's important to discuss them in ordinary times, because if an extraordinary time arises it's too late.
Having an entirely unchecked legislature, even if the existing checks are largely symbolic, is distinctly undesirable. That we have a unicameral legislature makes the GG's reserve powers all the more important. Note that it's a reserve power to refuse assent, not a reserve power to grant it. By definition, a reserve power is meant to be exercised sparingly.
On the chain-of-command issue, Commander-in-Chief > Minister of Defence. It's permissible, but highly irregular, for someone higher to give orders to someone who's not their immediate subordinate. If the CiC went around the Minister, that would be a theoretically legal exercise of command authority. Again, something that shouldn't be required in a functional democracy.
-
Oh, but America is Far Too Special And Unique for paper to work.
(See also: voting on a Saturday, having a tolerable health care system, etc)
I believe that voting on Tuesday is written into the Constitution, no? Changing that would be no small feat.
As for paper, the usual argument is that paper takes too long because America's got such a large population. Cue bleating from t3h meedja, who demand instant gratification and to know the result 30 seconds after the polls close. Who needs accuracy and verifiability in an election anyway? It's not as though they're voting for anyone important or anything, only the most powerful man in the world, with the ability to fuck up the entire global economy. Far more important to let those journos get to bed at a reasonable hour after telling the world who their next leader is.
Of course, as is observed regularly on Slashdot in stories about electronic voting, the larger the electorate the larger the pool of potential counters. It's a perfectly scalable system, and all the more so if you use OCR and only hand-count ballots that can't be read by machine. But that does make it much harder to do dodgy things to the result, since the ballots are readily available for recounts and verification. -
Yes, touch screens aren't infallible. Making sure that they're registering touches in the correct places, especially when the touched area is relatively small, is important. All the more so when you're voting for the leader of the world (not that most Merkins actually recognise how much impact their choice has on the rest of us).