|
|
|
|
|
'No' vote for state's electronic balloting |
By Mike Himowitz,
Baltimore Sun, September 28, 2006
[link
to article]
For proof of the adage that the road to hell is paved with good intentions, look no further than the superhighway that Maryland has constructed for the 2006 election.
So far, officials here have spent five years and committed $106 million to state-of-the art electronic voting.
Their intention: to prevent any repeat of Florida's 2000 presidential election debacle inside our borders. The result: a system so embroiled in technical problems and political squabbling that the governor who approved it is making robo-calls to voters advising them to vote by absentee ballot.
There's no question that the hardware and software Maryland bought from Diebold Election Systems Inc. failed one mission miserably Sept. 12, when computerized voter identification machines crashed during the primary and forced thousands to cast provisional ballots. Some were still being counted this week.
If Diebold hasn't fixed the problem by Nov. 7, the State Board of Elections will revert to old-fashioned paper voter lists. Which, come to think of it, isn't such an awful prospect.
After all, we've managed to get along with paper voter lists for the last two or three centuries. They don't crash. And that may be the real lesson of this election: throwing technology at a problem doesn't necessarily solve it.
For years, concerned computer scientists and a handful of ink-stained techno-wretches (like me) have been warning anyone who will listen that you can't trust all-electronic voting systems - or people who blindly believe in them.
One reason is that minor software glitches like the one that made Maryland's primary so "interesting" can easily disenfranchise tens of thousands of voters - who are already more suspicious than ever after the Florida fiasco of 2000.
Worse yet, no all-electronic system of counting votes is safe from glitch or gimmick. And thanks to Maryland's own efficiency, this danger now extends to every voter in the state, not only this year, but in future elections.
Why? Back in 2001, when officials contemplated replacing the hodgepodge of manual and electronic voting machines used by Maryland's 24 jurisdictions, they thought they would do the right thing: adopt a single, consistent system for the entire state.
Of course, that reasoning assumes that the system is secure and reliable. But what if it isn't?
With a universal hardware like the Diebold AccuVote-TS terminal that Maryland has adopted, a single mistake or well-hidden line of malicious programming code can corrupt an entire statewide election, instead of a handful of returns from a single jurisdiction.
Which brings us to the Board of Elections' second assumption: high-tech is better than low-tech.
When it chose all-electronic terminals, it rejected a hybrid system of scanned paper ballots that had already served voters well in Baltimore County and other jurisdictions for a decade or more.
These scanner-based systems aren't flashy and slick - and they have a big disadvantage if you're an election administrator. You have to print, store and distribute paper ballots. No question - that's a pain in the neck.
But scanner systems do have three advantages. They tabulate votes quickly and accurately. They require a minimal investment (one or two scanners per polling place instead of a dozen or more touch-screen terminals).
Most important, in case of disputes, the paper ballots are available to be scanned again - or even counted by hand. When all the votes are invisible, what do you recount?
Still not convinced? Still not convinced we've made a mistake with an all-electronic system? Fire up your Web browser and check out a video prepared by students of Edward W. Felten, director of Princeton University's Center for Information Technology Policy. You'll find it at http://itpolicy.princeton.edu/voting/
Felten is a gadfly on public technology issues and one of many prominent computer scientists who think we're crazy to trust our franchise to electronic systems without any physical, real-world backup.
The video demonstrates how Felten's students were able to break into a Diebold terminal with a key from a hotel mini-bar. That done, they inserted a memory card with a malicious virus that deliberately misrecords votes.
The virus spreads from one terminal to another by infecting any card inserted into its memory slot and installing itself on the next machine that card is plugged into. It's not as fast or efficient as a network virus, but it does the job.
Unlike attempts to rig old mechanical machines - which affected every vote cast - a sabotaged counting program can manipulate a close election by switching one out of every 40 or 50 votes.
Diebold response That can be hard to spot. And Felten's virus makes it almost impossible by erasing itself from memory the moment the election is over.
(You'll find Diebold's response at www.diebold.com/dieboldes/)
None of these criticisms is new. When Aviel Rubin, a Johns Hopkins University security specialist, first raised the issue in Maryland in 2003, state election officials painted him as a nut and rabble-rouser.
But in study after study, in state after state, he and other critics proved otherwise. Eventually, the National Science Foundation put Rubin in charge of a $7.5 million grant to come up with secure voting technology.
Meanwhile, Maryland's election system has become the epicenter of a tectonic squabble between Republican Gov. Robert L. Ehrlich Jr., a Democratic General Assembly and a split Board of Elections with an Ehrlich majority but a Democratic chairman, Linda H. Lamone, whom Ehrlich would love to get rid of - but can't.
Trailing Baltimore Mayor Martin O'Malley in the polls, the Guv has already intimated that he doesn't trust the system - even though he approved its purchase. And Lamone's answer to critics has never wavered. "Trust us," she says.
Well, I've said before: I don't trust them and you shouldn't either.
Just look what happened on Sept. 12. Under the board's steely-eyed supervision, Diebold unleashed E-poll software that had never been tested with a full day's simulated voting.
Tom Feehan, Diebold Election Systems' project manager, called that failure "an oversight." Well, it's not an oversight - it's a boneheaded stunt no rookie programmer would pull unless he wanted to become an unemployed programmer the next day. At best, it's pure nonfeasance.
Trust us? Yeah, sure. Maybe we should all take Ehrlich's suggestion and ask for paper absentee ballots if we want to make sure our votes are counted, because we're stuck with what we have for this election.
Whatever your politics, pray for results that aren't close enough for a few glitches - or even some rogue programming - to make the difference.
And when it's over, send these terminals off to some other state, or even to a recycler, and turn them into, I don't know - high-tech Etch-a-Sketches.
Then let's buy a system for future elections that we can verify and believe in.
|
|
| |
| |
|
|
