One of my read­ers (yes I actu­ally have at least one that’s not related or an out­side friend) pointed out an arti­cle that he was work­ing that has been picked up by the media.

The issue it seems is in the method that Ohio bal­lots are marked and recorded. While I will pro­vide you with the links to the tech­ni­cal side of this type of attack, the results are tragic in the face of democracy.

Our sys­tem of gov­ern­ment relies on a secret bal­lot. Some peo­ple have no prob­lem shar­ing who and what they are vot­ing for. Oth­ers how­ever feel undo pres­sure from peers if some­one else knows how they may vote on unpop­u­lar issues. For exam­ple, that prop­erty tax hike that you cam­paigned against for months and invested time and money in fight­ing, well you could go down and see all the peo­ple that voted for it after it past. Some less scrupu­lous peo­ple could use this infor­ma­tion to harass the vot­ers that passed it. After this harass­ment has taken place there will be more out­side forces that can cause that voter not to vote their con­science the next time.

Since this is one of those rare oppor­tu­ni­ties where you can ask the per­son that dis­cov­ered this bug (he e-mailed me first) I sent this back to him

“.….I’m just curi­ous what your thoughts are on how to fix this hole?

Do you wish them not record time stamps or not have a paper trail? Elec­tronic vot­ing machines have sev­eral known vul­ner­a­bil­i­ties that can lead them to manip­u­la­tion? I can see that you would rather have a bal­lot form of paper trail since these could the­o­ret­i­cally be mixed up so you can’t them orderly, but at the same time the anonymity is only going to be aggre­gate to the amount of data you col­lect. In other words in small elec­tions turn outs it will be eas­ier to deci­pher who voted which way then a large voter turn out.

An easy imme­di­ate step to help insure secu­rity of these records with­out the addi­tional cost of vot­ing machines would be have the poll­sters record the vot­ing num­bers ran­domly instead of orderly with­out time stamps since we only really need to ver­ify they are eli­gi­ble and do not need to record which order they came in to vote accurately.……”

A few hours later (we all have day jobs) he sent me this response:

“…I still see elec­tronic vot­ing as supe­rior to hand-counted bal­lots. My opin­ion on the details changes ;-) but cur­rently I would like to see imple­mented

  • sep­a­ra­tion of ‘vote’ and ‘event’ logs (sim­i­lar to ES&S’s Unity database)
  • sort­ing of ‘vote’ logs by con­tent, rather than a decep­tively inse­cure [pseudo]random sequence.…”

I have some per­sonal issues of course with elec­tronic vot­ing machines and while I’m not going to put in the com­plete exchange we dis­cussed here is the rel­e­vant part of my reply.

“.…..I’ll agree that elec­tronic vot­ing machine are supe­rior, espe­cially in the amount of time to tally the votes and accu­racy. The things I would like to see imple­mented is an open source approach to the code (I’m not a linux zealot as most my com­puter are win­dows) as this can have a ver­i­fied secu­rity through the process of code review. The main coders would just have to pub­lish the code but main­tain all changes by the in house pro­gram­mers after sug­ges­tions of a code review — since it would not be advis­able to take cus­tom code in from nor­mal coders — the main­tain­ers would still be account­able for accu­racy and reliability.

The next thing would be a split printer paper trail where it could main­tain one copy for archival pur­poses and pos­si­ble man­ual tal­ly­ing to ver­ify votes were counted accu­rately by the com­puter and a mir­rored copy that would be printed out for the voter to ver­ify the accu­racy of their vote. While there would need to be a mech­a­nism for the archival copy to be destroyed if the voter decided to change their vote or found an inac­cu­racy in their printed bal­lot — this for the most part would be a triv­ial propo­si­tion of an enhanced paper tray.

But all in all I agree with you — of course for Ohio vul­ner­a­bil­i­ties the intern that took the back up home on a flash drive full of pub­lic records as the states way of main­tain­ing off site records is a far worse expo­sure that I hope can be cor­rected along with the vot­ing issue.…..”

Now this is some­thing that really needs to be addressed quickly. It does not effect all Ohio coun­ties but please pres­sure your offi­cials to make sure that you are not using the type of equip­ment that can cause this expo­sure to your pri­vate data. In the end your pri­vacy is all you have that will truly always be yours.

Here is a link to Jim’s site

A link to the C|NET article

Ed Felton’s view at Free­dom to Tinker

blog comments powered by Disqus