Comment Re:Mercifully... (Score 1) 197
I'm hoping you can provide an example -- anywhere -- where you have had an opportunity to compare the paper against the machine. Of course, they did that in Cleveland, the only place in history I think that has ever actually compared the DRE "voter verifiable paper audit trail" with what the machines said. That cost Cleveland about $350,000.
It didn't match.
I am currently up to my ears in "auditing." First, you can't get the records. Those records you are allowed to look at, you can't do so in a timely manner, generally until after all recount and contest periods have expired. And those records you get to look at generally have information on them that doesn't reconcile. When it doesn't reconcile, you don't get meaningful answers. And there are no consequences for public officials who run mismatched elections, except in the rarest of cases, such as what's happening in Sarasota right now -- but that took citizens groups from 3 counties and a team of lawyers. Sarasota citizens take back their elections
In every election, we are seeing more votes than voters, zero reports that aren't zero, machines that don't match their paper results reports, lost votes, but mostly, we aren't allowed to see. A little-known secret is that even in locations where they have "random x% manual audits" the random is "selected" (you heard me) and the audits frequently don't match. When they don't, there is no expansion of the audit, and indeed, there is not even a disclosure to the public that it didn't match.
If you are persistent enough to find out when the "random" audit will take place, you MIGHT be able to watch, and if you are allowed to watch, you MIGHT be able to actually see anything. It is common for them to put you too far away to view the ballots, just as it is common on Election Night for them to turn the computer screens away from the observer area so that observers get to watch the back of the monitor, not the front.
Theories of how elections should work run into a bumpy ride when you watch first-hand how elections are actually run.
But let's suppose we solve those problems.
With computerized voting, it takes a small wheelbarrow and about 60 days free time to engage in stupefyingly tedious but precise work to audit just one jurisdiction. You've got to check the (Diebold) computerized voter registration system with the (Diebold) electronic pollbook and then look at the (Diebold) voting machine results tapes and then compare them with the (Diebold) central tabulator report. But all those things can be rigged, as we showed in the HBO film "Hacking Democracy." What remains is counting all the paper votes. Not a sample. All of them.
But counting the "VVPAT" is almost a comedy. It's printed on 8-pt type on a roll made of thermal paper that is about 3 inches wide and a few hundred feet long. If you thought staring at chads was bad, you'll croak when you see this. But that's not all.
The chain of custody for the paper trail, which you MIGHT get to see after weeks have passed, is another of life's little mysteries. But let's assume the chain of custody is acceptable. In San Diego, they charge a dollar a vote to look at the ballots. To look at just 11 precincts (out of over 500) citizens were hit with an unitemized bill for over $8000. At least they could look -- in Marin County, the race was dictated by the absentee ballots, and citizens were told that it is not possible to sort the absentees for just their candidate, so they'd have the choice of paying for looking at ALL of them or paying an even bigger fee to have them sorted.
But that's in the more voter-friendly locations. In Nebraska, a losing candidate for senate tried to purchase a hand recount. "There is no provision in the law" he was told, and I have that letter from the Nebraska Secretary of State, "for you to count the ballots by hand." They must go through the machine again.
Again, why are we doing this?
A bill has been introduced in congress, HB 6200, that finally has the right idea. HB 6200 would require all votes for the presidency in 2008 to be hand counted at the precinct, and precinct size would be restricted to no more than 500. Canada does this in federal races, with very little controversy. New Hampshire hand-counted in public at the polling place for 45% of its jurisdictions this year.
We threw technology at something that needed nothing more than a check-box on voter registrations (like the organ donor checkbox on a driver's licence). You check the box if you are willing to volunteer for a hand count team. Hand counting at the precinct is not a computer problem, it is a project management challenge, and there is excellent software for that. When you hand count, you don't have long lines because someone shorted the machines, they broke, the encoders are on the fritz, or someone forget to pack the keys to the memory card bays.
It's time to rethink this whole deal.
Instead of thinking up new toys, and toys to check the toys, invest the time and money into design and testing of ballots for maximum accuracy and speed in hand counting. In New Hampshire, they discovered that if you sort and stack before you count, accuracy goes up and speed goes up too.
But there's no market then for a certification industry, independent testing labs, committees to set standards, purchasing machines, upgrades, tech support, election media programming, or annual license fees.
And the computer experts who have been cashing in with grants and consulting fees would have no relevance.
It didn't match.
I am currently up to my ears in "auditing." First, you can't get the records. Those records you are allowed to look at, you can't do so in a timely manner, generally until after all recount and contest periods have expired. And those records you get to look at generally have information on them that doesn't reconcile. When it doesn't reconcile, you don't get meaningful answers. And there are no consequences for public officials who run mismatched elections, except in the rarest of cases, such as what's happening in Sarasota right now -- but that took citizens groups from 3 counties and a team of lawyers. Sarasota citizens take back their elections
In every election, we are seeing more votes than voters, zero reports that aren't zero, machines that don't match their paper results reports, lost votes, but mostly, we aren't allowed to see. A little-known secret is that even in locations where they have "random x% manual audits" the random is "selected" (you heard me) and the audits frequently don't match. When they don't, there is no expansion of the audit, and indeed, there is not even a disclosure to the public that it didn't match.
If you are persistent enough to find out when the "random" audit will take place, you MIGHT be able to watch, and if you are allowed to watch, you MIGHT be able to actually see anything. It is common for them to put you too far away to view the ballots, just as it is common on Election Night for them to turn the computer screens away from the observer area so that observers get to watch the back of the monitor, not the front.
Theories of how elections should work run into a bumpy ride when you watch first-hand how elections are actually run.
But let's suppose we solve those problems.
With computerized voting, it takes a small wheelbarrow and about 60 days free time to engage in stupefyingly tedious but precise work to audit just one jurisdiction. You've got to check the (Diebold) computerized voter registration system with the (Diebold) electronic pollbook and then look at the (Diebold) voting machine results tapes and then compare them with the (Diebold) central tabulator report. But all those things can be rigged, as we showed in the HBO film "Hacking Democracy." What remains is counting all the paper votes. Not a sample. All of them.
But counting the "VVPAT" is almost a comedy. It's printed on 8-pt type on a roll made of thermal paper that is about 3 inches wide and a few hundred feet long. If you thought staring at chads was bad, you'll croak when you see this. But that's not all.
The chain of custody for the paper trail, which you MIGHT get to see after weeks have passed, is another of life's little mysteries. But let's assume the chain of custody is acceptable. In San Diego, they charge a dollar a vote to look at the ballots. To look at just 11 precincts (out of over 500) citizens were hit with an unitemized bill for over $8000. At least they could look -- in Marin County, the race was dictated by the absentee ballots, and citizens were told that it is not possible to sort the absentees for just their candidate, so they'd have the choice of paying for looking at ALL of them or paying an even bigger fee to have them sorted.
But that's in the more voter-friendly locations. In Nebraska, a losing candidate for senate tried to purchase a hand recount. "There is no provision in the law" he was told, and I have that letter from the Nebraska Secretary of State, "for you to count the ballots by hand." They must go through the machine again.
Again, why are we doing this?
A bill has been introduced in congress, HB 6200, that finally has the right idea. HB 6200 would require all votes for the presidency in 2008 to be hand counted at the precinct, and precinct size would be restricted to no more than 500. Canada does this in federal races, with very little controversy. New Hampshire hand-counted in public at the polling place for 45% of its jurisdictions this year.
We threw technology at something that needed nothing more than a check-box on voter registrations (like the organ donor checkbox on a driver's licence). You check the box if you are willing to volunteer for a hand count team. Hand counting at the precinct is not a computer problem, it is a project management challenge, and there is excellent software for that. When you hand count, you don't have long lines because someone shorted the machines, they broke, the encoders are on the fritz, or someone forget to pack the keys to the memory card bays.
It's time to rethink this whole deal.
Instead of thinking up new toys, and toys to check the toys, invest the time and money into design and testing of ballots for maximum accuracy and speed in hand counting. In New Hampshire, they discovered that if you sort and stack before you count, accuracy goes up and speed goes up too.
But there's no market then for a certification industry, independent testing labs, committees to set standards, purchasing machines, upgrades, tech support, election media programming, or annual license fees.
And the computer experts who have been cashing in with grants and consulting fees would have no relevance.
