The California Secretary of State has finally released the source-code review portion of its two-month “top-to-bottom” examination of electronic voting systems certified for use in California, and it’s not pretty. “The software contains serious design flaws that have led directly to specific vulnerabilities that attackers could exploit to affect election outcomes,” the report concludes. “An attack could plausibly be accomplished by a single skilled individual with temporary access to a single voting machine. The damage could be extensive–malicious code could spread to every voting machine in polling places and to county election servers.”

And it gets worse. Princeton professor Ed Felten read through the Diebold report, as well as those of Hart InterCivic and Sequoia Voting Systems, and found that some of the problems it identifies are the same ones Diebold claimed to have fixed years ago. “Diebold claimed (p. 11) in 2003 that its use of hard-coded passwords was ‘resolved in subsequent versions of the software,’ ” Felten notes. “Yet the current version still uses at least two hard-coded passwords–one is “diebold” (report, p. 46) and another is the eight-byte sequence 1,2,3,4,5,6,7,8 (report, p. 45).”

Now, “1,2,3,4,5,6,7,8″ is an improvement over “11111,” Diebold’s last hard-coded security key, in that it employs eight numbers instead of just one. But surely it can’t be among those that inspired California Secretary of State Debra Bowen to recertify Diebold’s machines for use in the 2008 elections. Presumably, “come up with a less laughable password” was a condition of recertification.

The access panel door on a Diebold AccuVote-TS voting machine–the door that protects the memory card that stores the votes and is the main barrier to the injection of a virus–can be opened with a standard key that is widely available on the Internet. The exact same key is used widely in office furniture, electronic equipment, jukeboxes and hotel minibars.”

–Princeton professor Ed Felten

With the presidential primary approaching, Diebold Election Systems is finally developing a voter-verified paper trail–of bad press. Earlier this week, the company made headlines when a team of investigators found fundamental security vulnerabilities in its touchscreen voting machines (as well as those of rivals Sequoia Voting Systems and Hart InterCivic).

Now it’s back in the news again, thanks to another government-ordered study that found its optical-scanning machines to be flawed as well. According to a report released by Florida Secretary of State Kurt Browning, Diebold’s AccuVote OS optical-scan voting devices could compromise the upcoming presidential primary elections in which they’re to be used. The machine’s “memory card can be preprogrammed to redistribute votes cast for selected candidates on that terminal, including swapping the votes for two candidates,” the report explains. “The attack can be carried out with low probability of detection, assuming that audit with paper ballots are infrequent and that programmed cards are not detected before use.”

An unsettling revelation for anyone concerned about this whole idea of “election integrity.” But never fear, Diebold has vowed to patch the vulnerabilities identified in the report by the Aug. 17 deadline given it by the state. If it doesn’t, it risks decertification, which some would argue might not be a bad idea at this point. Remember, Diebold is the company that designed its widely criticized electronic-voting systems, to be opened with a hotel minibar key and then posted a detailed photograph of that key to its online store.

It’s the company that can’t seem to safeguard its source code. It’s the company that evaded election transparency laws in North Carolina. And it’s the company that modified its machines without notifying election officials. Twice.

Pretty much says it all, no? “What surprised me was not that we found things; what surprised me is the number of issues that we found,” Bishop said later in an interview with the Daily Democrat. “The problems covered the spectrum from physical locks all the way into software.”

Apparently, if one set out to design a voting system that prevents checks and balances, it would be hard to outdo the one used in California these days. Of course, the manufacturers of those systems would disagree with that assessment. As they have with Bowen’s review and the innumerable studies that preceded it. “While this evaluation was an interesting and helpful theoretical exercise, it did not represent a security-risk analysis and as such does not measure the severity of the actual threats in any meaningful way,” Sequoia said in a statement. “This was not a security-risk evaluation but an unrealistic worst-case scenario evaluation limited to malicious tests, studies and analysis performed in a laboratory environment by computer-security experts with unfettered access to the machines and software over several weeks.”

OK, so all electronic-voting systems are vulnerable in a laboratory environment. But have we forgotten that they’re also beset with problems in real-world environments as well? Remember the November 2006 elections? Not exactly smooth sailing, now was it? “Here’s the bottom line–no one trusts those machines anymore,” Markos “Kos” Moulitsas wrote at the time. “And not only do they damage the integrity of our democracy, but they give losing campaigns an excuse to grandstand and further erode faith in our system.”