- #71
plover
Homework Helper
- 191
- 2
This is a misreading. By confirming the vote, I mean the voter has some way to tell that the choices made are what become the permanent record. Systems as they stand are not secure enough for sufficient confidence in that without a paper record. I have no idea where the idea of the voter changing something after the vote is cast came from.russ_watters said:
In the most meaningful sense, it was the invention of electronic voting machines 30 years ago or whenever it was. Recent events have caused widespread adoption of still flawed systems which brings the problems to the fore. Computer security risks always increase with market penetration.
This is why what you're saying keeps sounding like conspiracy garbage. The problems were there before recent elections, and they were being solved. They became urgent problems when events caused an acceleration of adoption of these machines before satisfactory solutions were implemented.
No, you misunderstand his point. The records kept by an ATM, the customer receipt and, more importantly, the internal audit record, are entirely different than for a voting machine because the transaction is not anonymous for an ATM. This is one source of the technical difficulty in producing a secure electronic voting system.
As Jones said, providing the equivalent of a receipt is just the simplest technical solution to the problem for the moment. No one has said it is ideal.
The standards that have been established for ATMs are also vastly more stringent and robust than the current FEC/NASED standards for voting machines (and most machines for this election were not even checked against the current 2002 standards but previous 1990 standards).
Not with sufficiently specified and enforced record keeping—which ATMs have, and cash registers, as a rule (always?), do not.
I've been talking about the security of voting systems, not the recent election. Whether or not anything fraudulent happened in the recent election (and I have never argued that it has), this problem still exists and is meaningful. Is this off topic for this thread? I suppose, but in the absence of any compelling evidence of non-negligible software problems or of fraud from the recent election, I do think it is the important issue. Did I become interested in this because I had heard of the possibility of fraud in the recent election? Of course. But the technical issues themselves are entirely separate from what problems they may or may not have caused in a specific circumstance. For purposes of this discussion, it is you who appears obsessed with the recent election. I would suggest reading my comments through a lens other than disdain for those who disagree with you.
You've introduced so many strange assumptions as to my motives, and made so many points based on insufficient knowledge of the actual state of voting technology and of the relevant computer security framework, that I really have no idea what you think I'm arguing. It has seemed pointless to try to construct a technical précis as I have not felt like sorting through these misconceptions thoroughly enough to figure out what exactly you're missing. I have pointed to the assessments of the people whose job it is to deal with these issues. While I have no particular belief that my view of the situation is a complete one, or is accurate in all details, I have no reason to believe my sources unreliable. You have provided no backup to your viewpoint, and the only argument you have made from my sources was based on a misinterpretation. I singled out the documents I did because I thought most people would prefer to read a summary rather than a transcript of FEC testimony or a conference paper, not because they were the best or most detailed.
A lot of the time your responses seem to be directed at arguments that the problems with electronic voting are directly analogous to those with paper voting. (And to the extent that problems arise from election administration, this is true of course, but this is not what you were saying.) The most serious problems with the current electronic systems arise from the nature of computers and software development and have no real analogy to previous problems—the worst being the possiblity of untraceable changes made to the tally database, occurring either in the process of data input or after the event, and being the result of either buggy code (the more probable cause) or malice.
The incident in Ohio where a precinct with ~600 voters recorded ~4000 votes for Bush is a good example. There was no obvious relationship between the votes cast and the recorded tally; there was no reason to believe it was fraud, (and if it was it was stupid anyway since it was so easily detectable). But without the paper record of individual votes, there would probably have been no way to assign results for the precinct. As Schneier points out, an invalid paper ballot effects one vote, and statistically the overall effect of these is generally self-cancelling, but a software malfunction can occur at any place along the chain of data collection, affecting a greater or lesser number of votes depending on the location, and affecting them in an arbitrary, non-statistically neutral fashion.
The most obvious things that comes out of reading about this stuff is that the overall issue of balancing the needs of ballot security and ballot anonymity are complex on all levels, technical and administrative, that people have put a lot of thought into achieving these ends, and that the problems are not always intuitive.
