[ale] OT: Voting machines cracked in California

[Jonathan Rickman]

On 8/3/07, tom <tfreeman at intel.digichem.net> wrote:
Nicely said even if I don't quite agree with you.

Patching up a system adopted in haste with improperly understood
characteristics doesn't seem to be a good use of taxpayer resources to me.
I may be wrong, but backing out and trying again more slowly and along
several different avenues of approach would ultimately be more robust,
better tested, and therefor ultimately a better use of money.

In this forum, for the moment, we can thus agree to disagree.
_______________________________________________


I agree with you, in theory. But in situations like this you have to accept
the reality that the state isn't going to just box up all these systems,
store them in a warehouse somewhere and go buy a whole new system. That
said...

The appropriate way to solve the balloting problem is to find a human
readable format that can also be made machine readable, not the other way
around. For example, my company does the NFL Pro Bowl and NHL All Star
paper balloting. We use good old Scantron machines to tabulate the results
here and then ship the ballots back to the leagues so that they can audit
the results if they so choose. Granted, the number of ballots is a lot lower
than a national election (though it is comparable to a state election), but
we only use 2 scanners and 2 people for the process. It takes 2 or 3 days to
complete. Our error rate is zero because of the human readable format. If
the machine spits out a ballot as unreadable, the operator can just key in
the proper information or disqualify the ballot virtually at a glance. I can
see no reason to believe that this proces can't scale up and out to
accomplish the same thing at a national level.

--
Jonathan
-------------- next part --------------
An HTML attachment was scrubbed...