Slashdot Mirror
Re-Vote Likely After E-Vote Data Mishandling
davecb writes "A California judge is likely to order a Berkeley city initiative back on the ballot because of local officials' mishandling of electronic voting machine data. A recount was not possible because the city failed to share necessary voting records, a violation of election laws. In a preliminary ruling Thursday, Judge Winifred Smith of the Alameda County Superior Court indicated she would nullify the defeat of a medical marijuana proposal in Berkeley in 2004 and order the measure put back on the ballot in a later election."
This should have been done in 2000
yes, I know it would have been expensive.
The Kruger Dunning explains most post on
For God's Sake, Legalize it already.
The case points to the dangers of electronic voting systems, which make it harder to ensure fair elections, Luke said.
How about "make it relatively trivial to rig an election".
Your mistake is an issue of scale. It's relatively easy to slip in one or two false paper ballots. It may not even be that hard to make the machine a little more picky when it comes to checking punchouts on the democrat side of the ballot. But there's backups, paper backups, that get checked and confirmed, even if at a small ratio. Someone watching the pile of ballots go through the machine can find it odd that mostly left-leaning candidates get kicked out as incomplete ballots. Little things can be snuck through easier.
But electronic... that's what you want when you want to do BIG lies. Just off the top of my head from the last 2 POTUS elections... cards coming preloaded with thousands of votes. Systems designed so that if you left a busy machine collecting votes and forgot to empty it out, it would kick over at 16384 to -16383 (funny how that happened in left-leaning counties, eh?). Funny "glitches" (I hate that word when it comes to elections) that lost entire counties of votes. Concerns that the system might be undercounting Demos and overcounting Repubs. Software that made it exceedingly easy to switch your entire ballot to republican on the last page, without really telling you it was. Or software that just preselected your candidates for you.
Add too all that... NO paper trail... NO hard copy in your hand to confirm... NO audit trail to be checked to ensure fairness and honesty. Just trust the magic box will tell the other, main, magic box, the correct vote, hope for the best, and ignore the man behind the curtain promising Ohio to Bush. Also, ignore those pesky pollsters and statisticians, they don't actually know what they're doing.
Really, the 2000 Florida situation was unique, because a swing of a few votes either way made a huge difference. But at least ya'll could go back and CHECK. In '04 all you got was "here's the number, if you don't like it too bad". I'd rather have a few weeks of checking to make sure everythings fair, rather than an instant biased result with no appeal.
The scale of the flaws of electronic voting far outweigh the flaws of mechanical voting. With mechanical, a few votes can get screwed up. With electronic, a whole election can.
If I knew the wedgies I gave you back in 6th grade would have resulted in this . . . I might have taken a moments pause.
1) "Voting" Machine that prints out a combination human/optically readable ballot. Human verifies the human part says what they want it to say. We don't want observers confirming this, that's why privacy sleeves have been used for years.
2) "Sorting" Machine that sorts the ballots based on the optically readable ballot. Human flips through the stacks and verifies all of the human parts say the same thing for that race. Observers can confirm.
3) A dumb "Counting" Machine that counts a stack of ballots (without needing to know whose ballots are in it). Human puts the resulting number in the tally under the human readable name on the stack. Observers can confirm. Totals of the entire stack before sorting and after counting each race to confirm that nobody misplaced a stack of ballots.
At each step of the process, a very simple machine (low cost, minimum requirements for certification, etc) performs a single task (and hopefully it will perform it well). And following every machine step comes a step where humans can verify that the step was performed correctly. Since the individual machines don't contain any state about the election at all, voting machine malfunction cannot lose votes, and any malfunctioning piece of equipment can be replaced by any other piece that works. If standards are defined for each step of the process, then multiple companies can compete, driving down prices, and in the event a company is unable to provide sufficient numbers of voting machines, the remainder can be bought from other companies.
Furthermore, many of the tampering problems with paper ballots (whether cast electronically or not) can be taken care of with forethought and work. Ballot stuffing with leftover ballots (or duplicates, or casting the ballots people turn in as incorrect) can be stopped by issuing numbered ballots and invalidating the remaining or wrong ballots. Likewise, lost ballots would be known based on the gaps in numbers. Preventing this from identifying the voter (based on, say, their position in line relative to a planted observer) can be done by packaging the ballots in blocks of 100 or so, pre-randomized within that block. This way at the end of the day, only the unused ballots of open packages have to be invalidated, the remainder can be invalidated block-by-block (bigger blocks: more random and more to invalidate from an open package at the end of the day. smaller blocks: less random but less cleanup at the end).
If I have been able to see further than others, it is because I bought a pair of binoculars.