Talk:DRE voting machine

From Wikipedia, the free encyclopedia

You have new messages (last change).

I'm pretty terrified of the idea of unverifiable, unauditable elections so someone more pro-DRE's should go through and NPOV things. I tried my best to present the good side of DRE's Zenyu 16:51, Dec 16, 2004 (UTC)

Contents

[edit] Security

Any discussion of DRE security is incomplete without mentioning the Princeton virus that demonstrated changing vote counts on the version of Diebold DREs used in some 2000-2002 US elections. [1]

Limiting compromise to source code gives the misleading impression that keeping the source code secure (and not public) will keep the DRE secure. The condition ""If the security of the DRE source code software is compromised"" could be more completely rephrased as: ""If the DRE is compromised (such as through hacks inserted in the source code, replacement of any software component, loading a computer virus, remote access via infra-red, bluetouth, wifi, modem, etc. or direct access to the hardware)"". -- Dan Oetting 70.59.60.176 16:10, 18 September 2006 (UTC)

[edit] Point of clarification

I'm from Sydney, Australia. Australia DOES NOT use any DRE or other voting machines - in fact no where in the whole of Australia that I'm aware of. It could be there electronic tallying occurs to count the ballot papers, however manual recount is possible and occurs in some cases. The voting system is administered by a Federal body, the Australian Electoral Commission at http://www.aec.gov.au/, who ensure the process is accurate, fair and non-partisan. While there are occasional allegations of vote fraud, the system enjoys the respect of a large majority of Australians.

I believe you are mistaken. DRE's are used in Canberra and according to the FAQ at the ACT Electoral Commission, the software is audited and open to the public for examination, but is not required to keep a paper trail, much less a voter verified paper trail. The security is much better than in the US because the software is public, was audited for functionality, and there are mechanisms in place to make sure the software on the machine matches the software that has been audited. However the lack of a audit trail means someone could introduce an exploitable bug that could go undetected, and look innocent once detected. This is especially true because the Linux OS it runs on top of is complex and hence is likely to have some undetected bugs. --Zenyu 14:35, Dec 27, 2004 (UTC)

[edit] Proposed merge

Voting machine is really talking about DRE voting machines and not other types like lever machines or such. Since the Voting machine page is largely sub-standard, if there are any iotas of goodness, they should be merged here, no? -- Joebeone (Talk) 23:04, 22 March 2006 (UTC)

Voting machine is no longer talking about DRE voting machines exclusively. 68.50.103.212 20:43, 26 March 2006 (UTC)

[edit] Updates

  • Vendor's claims of security: updated for accuracy and obvious NPOV
  • The DRE voting machines typically take the form of a box or enclosure (rather like an ATM) or a laptop computer.
This is inaccurate and has been removed. If accurate it would still be incorrectly labled under Benefits of computerized voting machines
  • Indian voting machines use a two-piece system with a balloting unit presenting the voter with a button (momentary switch) for each choice connected by a cable to an electronic ballot box. [2][3]
Incorrectly labled under Benefits of computerized voting machines moved to Types of DRE voting machines
  • DRE voting systems are often favored because they can incorporate assistive technologies for handicapped people, allowing them to vote without involving another person in the process. They can also offer immediate feedback on the validity of a particular ballot so that the voter can have an opportunity to correct problems if they are noticed. This is particularly important, with a good user interface the problem of spoiled ballots could be greatly reduced by using computerized voting machines. If the votes are not counted however, there is little point in determining the intent of the voter.
Rewritten for clarity and accuracy.
  • Some precincts require that the entire ballot be presented to the voter simultaneously. This requirement was intended to reduce undervotes, when a voter votes in some races but misses others. DRE systems in these polling places need particularly large screens to accommodate all choices and it was discovered in the 2004 U.S. elections that this increased undervotes because the voter was required to push a small VOTE button on the screen before the votes would be counted, and many voters missed this and their vote was not counted.
This requires a source. Also, this would not be considered an undervote. An undervote would require a partial vote. If this information is accurate, none of the voters affected would have registered as voted at all. Lastly, why would thi be listed as a Benefits?
I could source this (people pushing a "vote" button too early or not at all) if we wanted to include it. The prose in italics would definitely need cleaning up (but it's probably best left off the page).
I believe the statement that "An undervote would require a partial vote." is incorrect. Undervotes are per each race/contest. They can be easily defined as "An undervote is where a voter votes for fewer than the minimum number of allowed choices for a given race or contest on a ballot." Many political and social scientists who study undervoting and overvoting use undervoting in top-ticket races as a proxy for a particular voting technology's rate of error. I think what you're referring to is a blank ballot being cast or the ballot not being cast at all because the voter forgot to push a "vote" button (as is characteristic to the Sequoia AVC Advantage machine and the Danaher ELECTronic 1242 machines). I suppose there could be a benefit argued for for full-face ballot systems although pagination means that it would be very easy to scale up in terms of ballot content.
I'm not arguing that any of this should go back, just being precise. -- Joebeone (Talk) 22:01, 26 March 2006 (UTC)
  • It is because of the benefits of DRE's and the HAVA that most DRE's have been installed, not because they make rigging elections easier. Some elections officials have even gone back to paper and pencil when informed of the problems with un-auditable DRE's.
Obviously this would need a source. The POV in this section is just not appropriate. ("not because they make rigging elections easier", come one). DRE's and HAVA section will be added.

Yikes. It looks like I wrote over some of your edits when I was trying to add <ref> tags. Hopefully, I put things back where they should have been. Sorry about that. Is it just me or does there seem to be something wrong with the MediaWiki software today? -- Joebeone (Talk) 22:28, 26 March 2006 (UTC)

Finally, I think I've got everything back between the accidental over-writes that happened... Here's a summary [4] of what's changed today. One problem I seem to have been having was that Wikipedia was serving up older versions of the page as the current version. Anyway, let's work from this if there are no issues. I know Rspeer valued some of the external links that were removed, we might consider adding them back. -- Joebeone (Talk) 22:46, 26 March 2006 (UTC)

[edit] External Links

What’s the general consensus here on what’s appropriate in the external links section?

  • A three year old Wired article?
  • Rebecca Mercuri’s opinion page (last updated 9/1/05)?
  • Links to groups that relate?

Personally I don’t see how these things are relevant. What do you think? Electiontechnology 09:12, 10 September 2006 (UTC)

[edit] VVPAT

I moved some information to the Voter Verified Paper Audit Trail page and added a "See" link to the main article. I also expanded it. I also added some more audit related information. I left he Dr. Mercuri info because it was directly related, but the crytography (Chaum, VoteHere...) were moved. I think they server a greater purpose. I really think the Voter Verified Paper Audit Trail is in need of an update. There has been a lot of new information relating to that topic recently. Electiontechnology 21:28, 26 October 2006 (UTC)

[edit] History

The article gives no indication of when DRE voting machines were first developed, where and when they were first deployed, etc. - Jmabel | Talk 19:52, 19 November 2006 (UTC)

[edit] NIST recommendations to eliminate DRE

I heard a news story yesterday to the effect that the N.I.S.T. had determined direct recording machines should be de-certified altogether, and that the voter verified paper trail (a roll of paper showing the votes) should also be eliminated in favor of a paper ballot (presumably printed by the machine and checked by the voter) due to numerous strange vote total events in recent elections. Does anyone have the official report for incorporation here? Edison 19:33, 30 November 2006 (UTC)

We're all awaiting the TGDC Plenary Meeting, December 4-5, 2006. Check out the prerelease materials. --Electiontechnology 19:42, 30 November 2006 (UTC)

[edit] Taintain's and Electiontechnology's edits

  • Auditing and IV is a fundamental challenge to all voting machines.
  • VVPAT is only possibly part of the process. Make it a subsection if you want, but there is much more to auditing and verification than the VVPAT.

--Electiontechnology 16:41, 13 February 2007 (UTC)

    • if You mean "recording votes" by "IV" it's no problem to an optical scan voting system. the cross is 100% where the voter makes it. why do You mention "auditing", that was not part of my edit. please stop reverting unless You give some reason for Your claims.
    • the article isn't clear about that now either, maybe You have a better idea to rework it.
    • lab attacks: i think there is enough emphasis on the lab environment, all these attacks could easily be done by an inside attacker. the SDU-attack could be done by anybody within 25m of the election office.
--Taintain 17:52, 13 February 2007 (UTC)
  • I mentioned "auditing" because it's a section about auditing. Maybe this is another one of your problems interpreting english...
  • If you don't feel that is clear, you are more than willing to edit for clarity, but reverting to inaccurate information is unacceptable.
  • Regardless if you think the attacks "could easily be done," they have not and implying otherwise is inaccurate.
  • "all voting machines" cannot incorporate assistive technologies.
  • You miss the point about disabled voters. It is not a matter of trust but anonymity.
  • DRE feedback is not similar to the feedback of optical scan voting systems. Just because they both provide feedback does not make them in anyway similar. It is inaccurate to say "similar to optical scan voting systems."

--Electiontechnology 00:03, 14 February 2007 (UTC)

    • please try to avoid the personal offense. i'm still missing any reason for You claiming that all voting machines have a challenge with their recorded votes. anyway i included i source for my claim, do You want me to roll out the concept of software independence in more detail or is the link enough ?
    • i wrote nowhere they could be easily done, i wrote they've been done in a lab environment any why. anyway i included a real life attack and wrote more details to the others. if i find the time i could make a section for each attack.
--Taintain 13:52, 14 February 2007 (UTC)
  • I'm not sure what you take offense to. I apologize if you think anything I said was intended as such. And to try to explain to you again, you seem to be missing the larger issue. This was not about software independence. It is about the entire process of a vote, from the voter through tabulation and including auditing. That is a fundamental challenge to all voting machines. I understand you take issue with one portion of that (the recording} and I even highlighted that in the following section. If you would like to put in a section on Software Independence, be my guest, but stop reverting accurate information for inaccurate information.
  • Maybe you have a short memory? You said "all these attacks could easily be done by an inside attacker" in you last bullet. All I'm saying is that it needs to be clear. Putting laboratory experiments under "Attacks on DREs" is misleading.

--Electiontechnology 16:44, 14 February 2007 (UTC)

    • to make it short: so what is the fundamental challenge of an optical scan system with the recording ? what's the use of pointing out the similarities with other systems, the differences matter.
    • i said that on the talk page, in the article i wrote "all have been done in laboratory", what You removed was "to avoid breaking any laws", why is that misleading ?
--Taintain 21:12, 15 February 2007 (UTC)
p.s.: please edit a bit more careful, You had one = too much and one section double
  • You in no way explained your edits. See my previous points and if you want to revert have a good reason.

--Electiontechnology 22:25, 15 February 2007 (UTC)

  • the is no fundamental challenge for an optical scan voting system in recording votes because it's just a paper ballot. so saying all voting machines have a challenge with recording is wrong. that's a good reason for reverting.
  • why did You delete the VVPAT-sentence ?

--Taintain 23:34, 15 February 2007 (UTC)

  • "recorded as cast" You have no idea what the optical scan machine recorded. I didn't say it was a challenge to record it. You misunderstand again. The very next sentence explains the difference with a voting system without a tangible ballot.
  • I'm sorry you are having trouble keeping up with our conversation. I already explained that. The VVPAT is one example of a system that "can include the ability for voters to verify how their votes were cast." The VVPAT is still there, but saying "can include the ability for voters to verify how their votes were cast, known as a voter verified paper audit trail (VVPAT)" is incorrect.

--Electiontechnology 23:53, 15 February 2007 (UTC) ==Caltech Study=] this is what the caltech study says:

Yet this report also suggests a cautionary tale about placing too much faith in buying new machines alone. The greatest residual vote rate gains were not made in the states where voting machines were upgraded in a piecemeal fashion. Rather, voting machine upgrades were the most effective when the entire state was involved. The best example was Georgia, which switched over entirely to one type of voting machine (Diebold Accuvote-TS’s) and engaged in an unprecedented voter education effort. It is likely that the combined effects of these two actions, not either one alone, that led to Georgia’s stunning improvements over time.

http://www.vote.caltech.edu/media/documents/vtp_wp21v2.3.pdf RESIDUAL VOTE IN THE 2004 ELECTION, p.15

and this was in the article:

A comparative analysis by the CalTech/MIT Voting Technology Project suggests that a switch to a DRE-based system results in a significant drop in the number of spoiled ballots. Most notable is the state of Georgia, which went from second-worst to second-best in the country in terms of its "residual voting" rate after changing its voting system statewide.

i removed it because i couldn't see any reasonable way to make that part useful for this article.

Retrieved from "http://en.wikipedia.org../../../d/r/e/Talk%7EDRE_voting_machine_7a7d.html"