pageok
pageok
pageok
How to Hack a Diebold Voting Machine:
Ariel Feldman, Alex Halderman, and Edward Felten have the scoop here, and even have a short video to go along with their paper.
JohnAnnArbor:
An ATM-style voting machine is like the old mechanical voting booths: who knows what happens when you hit the "enter" key or throw the handle?

Paper, machine-scannable ballots are the way to go.....
9.19.2006 9:58pm
Chris Bell (mail):
And John, what will you do when the ballot scanner is also rigged?

"Power resides not in he who votes but in he who counts the votes." ~ Joseph Stalin
9.19.2006 10:19pm
JoeHall (www):
That Stalin quote has never been referenced to him.
9.19.2006 10:30pm
JohnAnnArbor:

And John, what will you do when the ballot scanner is also rigged?

That's why God invented hand recounts, which the paper ballots make easy, but the other machines make impossible. I thought that was so blindingly obvious it didn't need to be explained.
9.19.2006 10:45pm
Chris Bell (mail):
But why have the hand recounts when the voting machines and the ballot scanners give the same result?
9.19.2006 11:36pm
billb:
Chris: Because you can when you don't trust Diebold.
9.19.2006 11:46pm
whig (www):
Videos here.
9.20.2006 2:47am
Lev:
Computer Science
and Technology--------------------------------------------NBS Special Publication 500-158, Accuracy, Integrity, and Security in Computerized Vote-Tallying,
Roy G. Saltman


...
>3.2.1 Vulnerabilities of Paper Ballots

>...Inaccurate counting: Hand counting of large numbers of paper ballots is generally inaccurate, because of human inattention and fatigue, compared with counting of machine-readable ballots....
9.20.2006 3:03am
Mary Katherine Day-Petrano (mail):
Another reason voting machines must be speech recognition assastive technology accessible (which the DIEBOLD machines are not), is it can produce both an electronic and paper trail if a hardcopy is printed out by the supervisor of elections, for doublechecking. In Florida, Fla. Stats., Secs. 282.601-282.604 would appear to require this.
9.20.2006 3:10am
Mary Katherine Day-Petrano (mail):
"assastive"=assistive
9.20.2006 3:11am
Gumbey (mail) (www):
I know a little about the diebold systems and voting in general. The voting machines are scary enough, but the process is even scarier. In Memphis, we have an election commission who appear to have taken a permanent leave of absence. Last year I uncovered evidence of people voting from vacant lots, felons voting, and ultimately deceased voters were "uncovered". I also found several thousand voters who were deceased, yet still on the voter rolls. There were several indictments, a state senator was unseated, etc.

In August, I ran our voter file against the TN driver's license file and found approximately 160 people who voted in the early voting, yet appear to live outside Shelby County. Today, I received a fresh copy of the entire 604k voters in Shelby County (Memphis, TN) and started checking for deceased voters. My program has been running for less than one hour and already there are 160 who appear to be deceased, yet are still on the rolls.

The bottom line is, the voting machines are problematic, but the process is even worse. In Memphis, a person need not produce any "real" identification. A voter can show up with a piece of paper that appears to be a bill, which has been sent to the address they list, and they are allowed to vote. I believe we should take peoples pictures and thumbprints when they vote. This would not be cost prohibitive and could be archived in the event an investigation were to be held.

Here are just a few names with dob that can be checked at the rootsweb ssdi site:

HARRISON LUMPKIN 18920506
LURA POPE 18940905
GEORGE WEAKS 18951120
PHIL SCHREIER 18970424
BESS SOLARI 18970703
RUTH BUSH 18970809
JOE CARTER 18971001

The link to the ssdi site is http://ssdi.rootsweb.com/cgi-bin/ssdi.cgi
You will need to click the advanced button to enter the dob.
9.20.2006 3:54am
Isaac (www):
Lev:

Hand recounts may have precision issues, but it is much more difficult to hack the process and introduce bias than it is when the process is a black box. If you repeat the hand recount multiple times, the average count is very likely to be very close to the actual number of votes, while no amount of recounting with a potentially corrupt black box will guarantee that the result is even near correct.
9.20.2006 1:33pm
Isaac (www):
In addition, the precision of a hand count can be (and probably has been) measured and expressed as a confidence interval, like "We are 95% sure that the hand count is correct to within 0.5% of the total count." This confidence interval is useful for deciding if the electronic totals are trustworthy: if the electronic total falls outside the confidence interval, you can say something like "We are 95% sure that the electronic results are incorrect." Then, you'll know that you have to toss the electronic results for this election and rely on the imprecise but unbiased hand recounts, and look into fixing the problem with the electronic system for future elections.
9.20.2006 1:43pm
Lev:
Isaac

The entire article is worth reading.


If you repeat the hand recount multiple times, the average count is very likely to be very close to the actual number of votes,


How many hand recounts are done to achieve that nirvana?


while no amount of recounting with a potentially corrupt black box will guarantee that the result is even near correct.


You either have a corrupt box, or you do not. If you do not, the box recount will be more correct than the hand recounts. If it is corrupt, then it may or may not be more correct than the hand recounts, which can also be corrupt.
9.21.2006 1:24am
Isaac (www):
Lev:

Did you read my second comment? One hand recount is enough to be useful in detecting a problem with the electronic results. Each additional count would have the effect of narrowing the confidence interval. There is no "nirvana," but there is the ability to make statistical statements about how sure we are of the results, which is absent in the case of a system that may or may not have bug-ridden or hacked code.


If it is corrupt, then it may or may not be more correct than the hand recounts, which can also be corrupt.


It is much more difficult to corrupt hand recounts than it is to corrupt software (Compare Sections 3.7.2 and 3.8.1 of the report you cited with Section 3.2.1). Many, many more people have to be in on it to get away with making a significant difference, as long as there are "effective administrative controls" (Section 3.2.1).
9.21.2006 10:10am
Lev:
Isaac

]One hand recount is enough to be useful in detecting a problem with the electronic results.

I see your point.

However:

]"effective administrative controls"

are still a problem. If you remember, in the 2000 Florida Extravaganza, it was discovered that a guy with access to the ballot boxes had in his possession a card punching machine with a whole bunch of blank ballots. If I remember corrrectly, that discovery was accidental.
9.22.2006 2:00am