New troubles reported with iPod-like e-voting units

With just over a week left before the general election, reports are surfacing of problems with e-voting machines from manufacturers rarely mentioned in the news.

Hart InterCivic isn't the most common e-voting platform out there (that would be the Accuvote TS and TSX line, from Premier / Diebold), and the eSlate doesn't operate like touchscreen units do. Instead, it uses a dial-and-button approach that's a bit like an older iPod -- run your finger around a big dial, then click a button to make your picks.

It's an interesting approach, especially since it's in theory friendlier to infirm hands than touchscreens have proven to be. (The eSlate's disability-friendly design was based in part on that of certain medical equipment.) But if you don't remember those pre-touchscreen iPods, the interface might prove confusing; in fact, tests earlier this year indicated that users found the eSlate interface particularly difficult to comprehend.

You can see a demo on the Hart Intercivic site, or perhaps Dan Wallach's walkthrough on accurate-voting.org will clarify matters. Wallach is the associate director of the ACCURATE project, and he was one of the computer scientists involved in the comprehensive California review of e-voting technologies in 2007, which recommended that the eSlate be decertified until the company addressed various security issues. (It did.)

The firm's problems haven't been limited to California, though; a VotersUnite.org survey of Hart InterCivic's news coverage turned up a number of problems around the nation (PDF available here), and an August report from VotersUnite detailed trouble with equipment used in the 2004 elections (PDF available here, see page 19). In addition, a whistle-blower suit was filed against the company in 2006 (PDF available here), alleging various improprieties connected to Hart InterCivic's pursuit of contracts in the early years of HAVA, the 2002 Help America Vote Act.

So now what's the matter? According to scattered early reports from Houston, some eSlate machines deployed around around the region have been acting weird, with instances of marking a ticket voted straight-Democratic as including a vote for Republican presidential candidate John McCain.

Though it's easy to dismiss at least some of the eSlate talk as what in another context would be called PEBKAC errors, the sheer logistics of putting November 4th together are boggling, and problems that seem trivial in isolation loom large when you multiply them by 50 states and less than 11 days. (One commenter on Wallach's own blog noted that the eSlate machines are a problem in his precinct because they don't ride in cars well and arrive at the polling places in need of care. These little things, they add up.)

And one more thing to watch, in light of recent hubbub over problems with voter-registration: A Houston Chronicle article concerning local voting glitches made mention of problems with the systems used to confirm voters' identification -- none of which have received the kind of scrutiny given over the past few years to the voting machines themselves. Just in case we haven't enough during the next eleven days to worry about.