Liminal states

{ 5 }

Comments

1. jon | 16-Feb-08 at 3:28 pm | Permalink

   And in New York, Sam Roberts reports in the Times

   Black voters are heavily represented in the 94th Election District in Harlem’s 70th Assembly District. Yet according to the unofficial results from the New York Democratic primary last week, not a single vote in the district was cast for Senator Barack Obama.

   That anomaly was not unique. In fact, a review by The New York Times of the unofficial results reported on primary night found about 80 election districts among the city’s 6,106 where Mr. Obama supposedly did not receive even one vote, including cases where he ran a respectable race in a nearby district.

   City election officials this week said that their formal review of the results, which will not be completed for weeks, had confirmed some major discrepancies between the vote totals reported publicly — and unofficially — on primary night and the actual tally on hundreds of voting machines across the city.

   Overall, Obama may pick up a few delegates.

2. jon | 28-Feb-08 at 5:25 pm | Permalink

   Eric Bidstrup’s Super Tuesday More trustworthy election systems via SDL? (Security Design Lifecycle) post on the Microsoft SWI Team’s blog looked at this from a computer security perspective. He looks at the security reviews of several voting systems, and highlights areas where software engineering best practices clearly had not been applied. My (belated) response:

   Good post, Eric, on an important topic.

   I was on the “Sufficient Evidence” National Academies/CSTB committee on Dependable Software with SDL co-author Steve Lipner a few years ago, and we had a workshop on election systems with David Dill, Avi Rubin, Douglass Jones, and Ted Selker — the workshop report is online on the National Academies’ site. It’s clear that there is there is plenty of room for improvement in the process and use of tools for election system software.

   One thing to highlight is that there are many important election security risks beyond the classics. I’ve discussed voting issues a series of threads on Liminal States like What democracy looks like in the US, February 2008. LA County suffered from a ballot design defect; design [in this sense] and usability reviews need to be considered from a security perspective as well. New Mexico had database quality issues after outsourcing maintenance; the full-lifecycle view of the SDL is vital here.

   And while it’s a contentious position, I’m of the belief that establishing public confidence in a full and fair election process requires open access to the software (including source code, a repeatable build process, instrumented binaries, and test harnesses) as well as documentation and test/inspection results. Trust has been shattered by current vendors; they — or the next generation — will need to be transparent to reestablish it.

3. jon | 10-Mar-08 at 4:14 pm | Permalink

   There’s now discussion about a vote-by-mail primary in Florida. Gulp. Rick Hasen is worried, too, in his Huffington Post column:

   But an all vote by mail primary makes me very nervous. Putting aside the fact that such a vote is not allowed under current Florida law and would need approval of the Florida legislature, vote by mail simply is not as secure as polling place voting. Vote by mail is essentially a mandatory absentee ballot election. Absentee balloting raises the specter of voter fraud and coercion, for the simple reason that polling officials are absent when voting choices are made. In the absence of a secret ballot, it becomes much easier to enter into an illegal vote buying contract, because the buyer can verify how the seller has voted. In addition, because voting takes place out of the public eye, the possibility of coercion or intimidation about how to vote becomes possible. Even if a Florida do-over would not produce a clear delegate winner between Clinton and Obama, it would have great political importance and could well influence the votes of the superdelegates, who will hold the balance of power if this issue goes to the convention.

   It might be that my concerns over the security of vote-by-mail in Florida are overstated. After all, Oregon’s vote by mail system has been touted as an excellent and fair system. Perhaps so for Oregon. But what is true for Oregon is not so true for Florida. One need only think back to the massive absentee ballot fraud in the 1997 Miami mayoral race that led a court to order a new election. And there’s something especially worrisome about rolling out a new system for counting votes for the first time in a presidential contest. It is like debuting your new play straight on Broadway.

4. Liminal states » Blog Archive » “Election falsification” and other voting issues in Ohio (updated) | 12-Mar-08 at 7:18 pm | Permalink

   […] the state’s pressuring Diebold to refund the $21 million for the decertified voting machines. Democracy in America, 2008. Nothing to see here, move along, move along […]

5. jon | 18-Mar-08 at 4:04 pm | Permalink

   Cory Doctorow’s Sequoia Voting Systems threatens Felten’s Princeton security research team on boingboing:

   Edwin Smith, the VP for “Compliance/Quality/Certification” at voting-machine manufacturer Sequoia Voting Systems has sent a threatening legal letter to Ed Felten — the Princeton law professor who’s led many security audits of voting machines in the past.

   The letter warns that if Felten and his colleagues publish any kind of security audit information of Sequoia’s machines (“Sequoia software, its behavior, reports regarding same”) that Sequoia will “take appropriate steps” through its “retained counsel.”

   It’s hard to imagine a stupider legal threat. Honestly.

   See discussions on Ed Felten’s Freedom to Tinker and Slashdot — and no doubt elsewhere.