Joseph_Daniel_Zukige writes
"I'm still trying to figure out who is doing what here. It looks like the typical bureaucratic mess, but it looks like NIST, operating under the Help America Vote Act has set up a Technical Guidelines Development Committee to advise the 'independent bipartisan' United States Election Assistance Commission. So, the TGDC is going to hold some public hearings, and they've invited members of the public to help them out: 'One hour will be reserved at the conclusion of each day for members of the public to provide up to five minutes of testimony.'" Read more below, including how to register (today is the deadline) for the meetings, which will take place in central Maryland later this month.
Update: 09/15 18:04 GMT by
T : Irvu writes
"You can submit online comments to NIST's Technical Guidelines process. The link is here. Just click on the link marked 'Submit Comments or Position Statements.' Alternately you can e-mail your comments to vote@nist.gov."
Joseph_Daniel_Zukige continues "I can't make it. (Very long drive across a very deep ocean, or plane tickets I can't afford.) Twelve people per session is not going to allow a lot of people to testify. I'm sure Microsoft has someone going to sell a MSWxx based voting machine. I hope somebody from the EFF is going. Think it would be possible to pack this thing with enough Slashdot geeks to convince the government at least that electronic voting absolutely requires a human-readable ballot to be produced?" The meetings are taking place on the 20th through 22nd of this month; you have only until 5 p.m. today to register, though.
From the linked PDF: "The meetings will be held at the National Institute of Standards and Technology North Campus, 820 West Diamond Avenue, Room 152, Gaithersburg, MD."
Slashdot Mirror
- Publicly accessable source - Whatever licening terms you use, the source code must be easily available to anyone and everyone who wants to read it.
- Verifiable binaries - Election inspectors must be able to verify the binaries installed on the machine by generating an MD5 (or equivelent) hashcode and comparing it to the published source.
- Paper Trail - The voting machine must keep a human-readble printed record of every vote cast. This is the only meaningful way to do recounts. In case of a discrepancy, the paper record should act as the real ballot...the electronic vote is just a fancy method of counting.
One thing I am opposed to is a "voting receipt" that the voter gets to confirm that thier vote has been cast. While this sounds good in theory, it's too easy for powerful organization (unions, corportations, etc...) to sway elections by paying people for voting by having them turn in thier voting receipts after an election.Canada's election consultants are doing their best to replace paper ballots with electronic systems.
The Chief Election Officer of Ontario has issued a report in favour of trying out electronic and Internet voting, and has already issued an RFQ for a technology pilot project.
You can read more about it on my blog Paper Vote Canada
We'll formalize this later but in "rough draft" form, here's our recommendations:
1) Open source. Not necessarily GNU licensed, but the source code of all voting systems must be publicly available on the vendor's website plus at least one gov't website if not multiple - choices include the county elections department's websites, the Federal Election Commission, state SecState sites, etc. ALONG WITH the compiler and operating system makes and versions under which the code was compiled; that will allow us geeks to do our own compiles and generate our own hash results so that we can compare with "in the field" binaries. (I have to disagree with Dr. Dent on his point #2 in that I don't want to have to trust somebody else's hash numbers...I want to roll my own.)
2) Voter verifiable paper trails. The best such schemes are similar to the one Avante developed - your vote is printed on a paper strip "behind glass". You get to look at it, make sure it's OK and if you like it, hit "OK" on the touchscreen. A "robot snipper" clips off that piece of paper, it drops to the bottom of a sealed bucket and it's the official vote of record in case of recount. You don't use a take-up reel because then you can cross-ref the voter order with the vote order and figure out who voted for what. The voter cannot later prove who they voted for (it's not a "reciept") - that way "Guido" can't breaka you legga for voting "wrong" or pay you for voting "right". Oh, and the paper vote of record has an encrypted bar code strip to ID false "extra bits of paper", and minor mistakes in the dot-matrix print that are hard to spot but form their own second tamper-code.
3) This is the major piece that Bev Harris has contributed. Harris used to be a forensic accountant, meaning she dug into financial fraud for a living. In any accounting system, there are auditing procedures and steps at EVERY step of the way as cash is handled. Votes need to be handled the same way - there's documentation every time they change hands, there's a REAL audit trail, and similar steps that need to come from the CPA community. As one example: in a real audit trail, if data entry was done wrong and needs to come out, it isn't erased. It's MARKED (and datestamped) as "not valid" but it's still in there so you can see what happened. None of the current systems do this, with the possible exception of Avante (I'd have to take another look on that point.) Diebold, Sequoia, ES&S and Hart sure don't!
4) Mandate Read-Only-Memory storage of votes at the terminals! This is another thing Avante got right - and no, they ain't paying me or BBV.org a red cent. Their voting terminals burn the vote data to CD-ROM. Diebold, ES&S and Sequoia burn data to PCMCIA memory cards...which can be stuck in a laptop, encryption cracked and the data messed with as happened in Volusia County FL, Nov2000.
---------------
This is PRELIMINARY and should be viewed as such, but it's a pretty good guide to where our heads are at. Blackboxvoting.org (not just a website, we're a non-profit public interest educational/research foundation) will be meeting to discuss a formal proposal ASAP.
Jim March
Member of the BBV.org board of directors (Bev Harris is our Executive Director)
I'm also a co-plaintiff (with Bev Harris) in the current lawsuit against Diebold in California which State Attorney General Bill Lockyer just joined.