Re: testing DREs - what a mess

[Paul Walmsley <paul@xxxxxxxxxxx>]

On Wed, 3 Dec 2003, Evan Daniel Ravitz wrote:

> On Wed, 3 Dec 2003, Neal McBurnett wrote:
> >
> > So this is another reason to avoid DREs and vote recording machines.
> > With a ballot marking machine, this sort of test is far less important
> > since the voter verifies the paper results on the spot.
> 
> You can verify the paper reflects your votes; but you can't verify
> that the scanner/computer combo counts them. And under the law, ONLY
> the scanner/computer combo is allowed to count them. The paper is
> false "verification" unless a judge ORDERS that it be counted by
> people instead of the machine. To my knowledge this has never or
> almost never happened in 30 years across the country.

Evan, 

What do you think about proposals to count a statistically-significant
percentage of the computer-scanned ballots by hand, and to compare the 
human assessment of those votes to the scanner's assessment of those 
votes?


- Paul