Grrr.

Both Lonewacko and the editor are ignorant about voting.

Voting is a very hard problem. As simple as it sounds, the secrecy issue makes it very tough. Then add it the real life issues of training poll workers, setting up procedures, 50 different states with 50 different rule sets, and you’re going to have problems.

It drives me nuts when people who are knowledgeable about technology think they are knowledgeable about voting.

The Caltech-MIT group are extremely neutral. It was they who did the study that showed that human-counted votes could be as accurate as machine counted votes, which was a surprising conclusion.

Yes, they seem to have an e-voting bias, but e-voting is the only system that can actually be improved, all other voting systems are static. So ultimately, they are the only voting systems worth discussion really. They do not have a bias for DRE systems necessarily, they’re just measuring the actual results of those systems. The way the wind is blowing is probably favoring DREs, whether there is a paper audit trail is another issue, and the audit trail need not be paper.

The VVAAT report? That’s a brainstorm about how an audit trail could be added to existing systems. LIGHTEN UP. As it points out: “The debate about verifiable voter audit trails has prematurely narrowed into two camps”

That’s very true. The “no paper necessary” crowd hasn’t yet presented a system that’s foolproof, while the paper crowd runs into other issues, as paper itself is no guarantee, it has to fit into the whole system.

Audit trail is not the same thing as paper trail…

Posted by: Opinionated Bastard at November 12, 2004 10:02 AM

..hey wacko..whatcha think.. [this is the link — LW]

Posted by: Rob_NC at November 12, 2004 10:15 AM

Both Lonewacko and the editor are ignorant about voting.

The “Editorial content” above is coming from Lonewacko.

If the Caltech-MIT group were completely unbiased, I would expect to see reports about how hackers could change votes, and about how current eVoting is full of holes.

I also put the link in the last comment into HTML format.

Here’s an idea, and perhaps “Opinionated Bastard” or someone else can tell me if this has a name and is currently used.

It’s a two step process.

The voter first fills out a normal paper ballot using some standard technique: filling in circles, connecting arrows, etc.

Then they walk over to a scanner with a monitor and insert the ballot into the scanner. The monitor displays their selections. If the voter sees everything is OK, they press “OK” and the ballot is ejected from the machine. The voter then puts that ballot into an envelope and puts that envelope into the ballot box.

The scanner records a preliminary count which stands as the official count unless something unexpected happens. A memory card in the scanner is physically transported to a central tabulator and the votes are read and published.

However, random samples of precincts are done to ensure that the machines worked OK and the data on the memory card matches the votes on the physical ballots.

In the case of a recount, the physical ballots are counted and they take precedence over the electronic count.

If the voter says the scanner’s output doesn’t match how they voted, that ballot is placed in a special bin, perhaps together with a note about the specific discrepancy. Then, the voter is given a completely new ballot and has to go through the process from the beginning.

If enough voters notice discrepancies, then we have first warning of a problem with that scanner or the ballot or other things.

I’d like to know if this system is used already, and if someone can point out flaws.

Posted by: Lonewacko at November 12, 2004 02:13 PM