Re: [Fwd: RE: MTBF for voting machines] part 2
Title: Re: [Fwd: RE: MTBF for voting machines] part 2
Actually, even the so-called "random" failures can introduce
skew. For example, some states have the custom of listing
the party of the governor on the "first" side (say left) of the
ballot, hence a misaligned touchscreen that shifts presses
from left to right, would always favor a particular party.
One can come up with numerous other scenarios where
random failures are not actually random when it comes to
voting applications, because of the machine and ballot designs.
Again, because there is no actual way to audit the intention
of the voters in a fully-electronic DRE, there is no way to
recover from losses, and since we insist that "every vote
counts" (even when this is known not to be true) and
elections can be (and are) decided on a single vote, the
random losses must be addressed.
Rebecca Mercuri.