MD Judge Gets E-voting Case Wrong -- Emergency Appeal Set for September 14
A judge in Maryland on Tuesday issued a ruling [PDF] denying a request that Maryland voters be offered the option to vote on paper in the November election and that election officials implement basic security measures sought by the state's own experts. It's a very bad decision, and the court had to ignore two esteemed security experts and embrace some pretty outrageous falsehoods to do it.
While the decision was not unexpected -- the judge dragged his feet about even deciding the case at all, and we expected that it would have to be appealed to a higher court -- it's still a disappointment. The appeal hearing, before the Maryland Supreme Court (called the Court of Appeals there) is already set for September 14. And if that court is unwilling to take a stand on this issue, the legislative session in MD will start again in January, and the court has sugested that perhaps the legislature should "review" the situation.
Read on below for more details:
First , a few silver linings:
1. The hearing pushed the state to adopt parallel testing, something it had been resisting. This is when some machines are pulled out of service on election day and tested to ensure that no trojan horse or other hidden process has been triggered. This is one of the safeguards that the plaintiffs, as well as Maryland's own experts, had long been demanding.
2. The judge emphasized that the state will now implement some of the "reasonable" demands security advocates made in the wake of three different investigations into the Maryland Diebold machines, something that the state had been slow to do prior to the hearing. The new security measures include encrypting the information as it passes between machines, changing the passwords (they were previously all the same!), and taking steps to ensure external security of the machines. Again, all of these should have been done a long time ago, but it does appear that pressure from the court case led the state finally to move on them, and officials will have to face the judge if they fail to do so now.
But here's the bad news: the court adopted some strikingly wrong factual findings, and reportedly prevented the plaintiffs from presenting tons of evidence to counter them. The wrong findings include:
1. That "all experts agree" that DREs are "much more secure and less vulnterable than paper ballots, even the optoscan ballots." This is just not true. The expert reports generally indicate that optical scan systems are the most secure. Moreover, as the leader of the RABA report Michael Wertheimer testified, regardless of what future systems could do, the current Diebold system gets an "F" in security.
2. The court says that "the votes have been counted accurately." Apparently the court didn't see fit to review EFF's amicus brief that presented more than 20 instances when that was not the case.
3. The court says "recounts have occurred with complete accuracy," a stunning statement given that DREs leave nothing to recount. They simply reprint the results that the machine tabulated in the first instance. Computers definitely can and do print out the same results twice. That's not the same thing as a recount under any reasonable definition of the word. Recounts are a second "counting" of the voters' intentions and that's not possible with DREs.