franchisers over enfranchisement
adn.com | alaska : State rebuffs raw vote demand
The state Division of Elections has refused to turn over its electronic voting files to the Democrats, arguing that the data format belongs to a private company and can’t be made public.
Of course, there is some technical wrangling going on that obscures the issue. What the Alaska state Democractic Party requested is the raw database of vote tallies from the 2004 state elections. Diebold claims that the format of the data is a trade secret, despite the fact that there are copies of Diebold databases available on the Internet. This in and of itself is not outrageous; I’m no lawyer, but I believe Diebold is required to defend its IP, regardless of whether or not it has been pirated.
However, if that is their defense, then their offer of an Excel spreadsheet undermines it, unless they are willing to admit to obscuring some of the original data. The ’secrets,’ for whatever they are worth, to a database format are simply the names, datatypes, and storage sizes of each column. Producing an Excel spreadsheet from the original data would provide every single piece of that information. In order to produce a spreadsheet that does not, they would need to either exclude columns from the database entirely, or merge two or more of them together. In either case that would tamper with the original vote data. So: What is their motivation? Do they really believe that an Excel spreadsheet is any less revealing than a .sql dump?
More importantly, what if they’re right? What if there’s some sort of database trigger or embedded code that is proprietary, and they really are incapable of reporting the exact original data? Our situation is then that property rights are primary, even in the face of credible allegations of widespread federal election fraud. Why is that okay?
In related news, Diebold’s founding CEO resigned in December for unspecified reasons, while the company blames Hurricane Katrina for their recent woes.
Related Posts:
- how much information content in our elections? (November, 2005)
- to whom does our identifying data belong? (October, 2005)
- Oops! I Misplaced 19,000,000 Acres. (October, 2005)
- Where are They Now? (January, 2008)
- Gay candidates penetrate state legislatures (May, 2006)
January 25th, 2006 at 2:23 pm
To be. . . (well, not fair, but perhaps accurate), an Excel spreadsheet might be used to completely denormalize the data. A lot of optimizations in the application layer might be guessed at if you had access to the database schema.
But I’m taking away a larger point: It’s absurd to have proprietary systems form the apparatus by which we democratically elect officials. Voting is the last defense against abuses of power.
Back to open-source voting software: what’s the status of these projects? I’m just starting to look into it. . .
January 26th, 2006 at 2:29 pm
I did some more in-depth searching on that:
So far http://evm2003.sourceforge.net/ is the only one I know of that is completely ‘free as in *’ (and in python omg)
http://www.votehere.com/ is a commercial venture that has opened their source… kinda. there is reference code available for inspection, and they have apparently made their code available to auditors, but there are parts that remain secret, mostly the heavy-crypto… which is concerning. i haven’t taken an in-depth look at the provided source yet (http://www.votehere.com/downloads.php). will let you know if there are any interesting omissions.