Florida's new electronic voting machines work just fine . . .
. . . provided that you turn them on.
Now we learn that a voting machine in Hillsborough County, used for the two-week early voting period before the Aug. 31 primary, was not counted in the results on election night. The machine had been left in a test mode and so it reported zero official votes.
That means the votes of 245 citizens who were dedicated enough to take time out of their day, go to the early voting polling place and cast a ballot, didn't initially get counted.
This is the second problem we know about arising from the Aug. 31 primary under the county's elections supervisor, Buddy Johnson. He was appointed in 2003 by Gov. Jeb Bush after the previous supervisor, Pam Iorio, left to run for mayor of Tampa.
The first problem was a preventable computer issue that held up voting results until after 5 a.m. As he explained the delay, Johnson declared it was better to get accurate results than fast results.
It turns out the results were neither fast, nor entirely accurate.
Oh, for sure, 245 missing votes is not many, out of roughly 120,000 votes cast in the county. Johnson made it a point to say that even after the missing votes were added back in, no election outcome was changed.
No blood, no foul, right?
First of all, not to beat a dead horse, but every citizen has the right to expect his or her vote to count. Almost is not good enough.
Second, you might recall that we had a close election here in Florida about four years ago _ decided by only 537 votes statewide.
So losing 245 votes in a single Florida county is a big deal.
But third, the fact that this omission could happen at all _ the fact that there was a system in place in which this kind of oversight was possible _ is astonishing.
It's not like, you know, we've had a lack of scrutiny of our elections these past four years.
It's not like the ENTIRE WORLD isn't watching and this isn't THE MOSTWORRIED-ABOUT PRESIDENTIAL ELECTION IN HISTORY or anything like that.
Furthermore, this was a machine used for early voting. It was set up two weeks in advance. It was used every day. For two weeks. In test mode.
Then came election night. They brought the machine's results cartridge in for its results to be counted.
And the machine reported zero votes.
I talked to Buddy Johnson on Monday. He was frank about what happened. The machine was left in test mode; his procedures had not contemplated such a possibility.
"It would not leap out at you," he said, "that this was in test mode." The printout the machine produces indicated a test, Johnson said. But the poll workers had not been trained to look for it.
I asked Johnson: Wasn't there any protocol in place to make sure the test of each machine had been completed? "That has not been done heretofore," he said.
Now new procedures will be in place. Each machine will have an inspection sticker to be affixed after its test is finished and closed. Poll workers, upon opening the machines for voting, will double-check that the printout tape and the computer screen both show the machine is open for official voting.
Pinellas County, which uses the same Sequoia voting machines as Hillsborough, verifies that each machine's test is completed and it is ready for voting. Pinellas also watches on election night for machines that report zero votes and double-checks them.
As I've said before, the machines themselves work fine. I do not believe in bogeymen conspiracies that they are rigged. Neither do the machines by themselves "lose" or "eat" votes when properly operated.
But human error, that's a problem. It would be a problem whether the machine were electronic, lever, punch-card or optical scan. No system can overcome human error.
What is the answer to human error? The only answer is poll worker training, and more training, and so many double-checks and redundancies and checklists as to be a royal pain in the neck. The supervisor of elections has exactly one job; it does not seem too much to ask that he or she get it right.