Electronic voting machines have been in the news a lot lately, with the general trend of amazingly absurd problems. Why is this so difficult? Is it just that a big government is consulting big companies that are trying to hit big revenues? Or is it because small counties are trying to fund system development and can't afford the talent to design it properly?
Given the reported failures of these eVoting machines (and the vote-count issue in Florida during the last presidential election) there is a natural reason to not trust vote counting systems. And why should we? Our right to vote is probably the single most important right we have, and if there is a reasonable doubt that our vote won't be counted, faith in the entire system goes away.
Since the Diebold's of the world haven't been able to get this right, I suggest we change the approach to what it should have been to begin with: The U.S. Government should fund (which means the people will fund) a project where the final goal is an open-source, openly documented, verifiable reference design.
In addition, I can't over-emphasize the need for proper User Interface and interaction design on a system like this. The audience is massive—probably more massive then for any other system design. It is everyone. And it's not like designing highway and traffic light systems. Sure, most everyone sees those, but most are weak designs that we are forced to obey either for our own safety or to avoid prosecution. A voting system needs to work and it's operation must be intuitive and error-free.
Once the final hardware/software spec is complete and publicly reviewed, the projected price per unit and number of units needed will be published. At this point, any manufacturer can step up and offer a product that conforms to the specifications.
This is a public voting system. Its design should belong to the public; and as voters we should be demanding the rights to see and verify the inner workings of the system.
[Updated]
On a related note, you might check out the Open Vote project if this topic is of interest. Cheers!