Sequoia Voting Systems had a problem in their tabulator that cropped up a couple times -- more notoriously in Albuquerque in 2002 -- that had a cap on how many files it would read, so it just lost more than 10,000 votes!
the good thing is that, in addition to the totals, there were records of the votes. So the vote records could still be tabulated. the fear is that there could be mistakes in both recording and tabulating (accident or fraud) and then there would be no way to know.
| || |
___Tech Policy/Security E-letter___ Written by washingtonpost.com's tech policy team, the e-mail version of this weekly feature includes an original news article and links to policy and cyber-security stories from the previous week.
Click Here for Free Sign-up
Read E-letter Archive
Sequoia debuted touchscreens with cartridge printers in Nevada this fall. They seemed to work pretty well. they provided the touchscreen experience with a paper backup. It might be the future.
It sounds like you are either on the payroll for Diebold or Karl Rove, and you sure are not a journalist. It is refreshing to see the rise of the IndyMedia movement, where reporters search for the truth, wherever it may lie, rather than your fake objectivism. Can't wait to see your face when your boy Bush loses on Tuesday, loses, that is, unless the pre-programmed Diebold machines perform as designed.
Knowing you won't post this from Alexandria
Dan Keating: OK.
I recently heard about people locally having trouble when voting a straight-party ticket on an e-voting machine. Apparently doing so makes it really easy to "skip" voting on local initiatives, etc., unless you make a special effort.
Now, granted, this was from "a friend of a friend." But it was a real, live friend of a real, live friend. Not an internet rumor. And it wasn't couched in partisan terms at all.
I'm not necessarily asking for comments on this specific case. But does this sound plausible to you in general? Are people addressing this issue?
Dan Keating: In 2000 and again in 2002, Texas had problems with straight-party votes -- on optical paper machines. The machine tabulators were mis-programmed. The good thing with misprogrammed paper ballot (or punchcard) counters is that they can be reprogrammed and the ballots can go back through. A misprogrammed electronic machine leaves no such fallback.
On straight-party electronic, some people have noted that different machines behave different ways. On some machines, you can push straight-party, but if you then switch any vote, all of the straight party is removed! On others, you can go straight-party and then make a few individual changes and it only changes the ones you actively alter.
Varying interfaces and behaviors with touchscreens are a big issue for research and recommendations.
Among the disturbing reports I've heard about electronic voting are some instances where people have shown up to vote, but apparently haven't. There was something like 140 blank votes for a primary in Florida and the excuse offered was that these people didn't realize their party didn't have a primary. That I doubt.
Now I've heard of the same thing happening in Prince George's County, MD and the excuse offered was that they must have been "protest votes."
It doesn't pass the smell test to me.
Can you comment on how this could happen and what the solution might be?
Dan Keating: I know i've already been accused of being in the bag for the industry, but I'll just want to point out . . .
In that famous florida case, there was only one race on the ballot. And something like 140 were left blank. Why would someone go in for a one-question ballot and not mark it? when that 140 blank was, I'm told by a researcher at MIT, less than 1% of the ballots, and some people like to have a perfect voting attendance record.
I agree that "protest votes" has been used as a smokescreen. When we reviewed the ballots in Florida, we found one-third of undervotes had clear voter intent. So there were some people who didn't want to make a choice and many who tried but failed.
since electronic is new to many people and also to pollworkers, it's not surprising that some people would fail to vote correctly. it should get better over time.
I understand that the machines are not networked together. However, why can't the individual machines be programmed to keep running tallies individually at certain periods during election day? I guess my question really concerns backing up the system. For instance, what if there are four terminals at my polling place, and I vote on Terminal A at noon on election day. Then, ten minutes after I leave, Terminal A crashes and the data on that terminal is not recoverable? If it's not networked to anything, is my vote lost?
Dan Keating: The terminal writes a record of each vote to the hard drive. So even if the machine crashes, the individual vote records are still there.
Hi Dan and Robert,
I live in Palm Beach County, Florida, the land of the hanging and pregnant chad that will be no more.
Can you truthfully tell me that if we have a recount that the voting machine on which I will exercise my consitutional responsibilities will accurate reflect my vote, that the machine will not crash, that the software will not go haywire, or the memory be unreadable?
....and without any redundancy or audit trail!;
Dan Keating: No one can guarantee you that. We can guarantee that you won't leave a hanging chad or overvote.