March 19, 2024

Interesting Email from Sequoia

A copy of an email I received has been passed around on various mailing lists. Several people, including reporters, have asked me to confirm its authenticity. Since everyone seems to have read it already, I might as well publish it here. Yes, it is genuine.

====

Sender: Smith, Ed [address redacted]@sequoiavote.com
To: ,
Subject: Sequoia Advantage voting machines from New Jersey
Date: Fri, Mar 14, 2008 at 6:16 PM

Dear Professors Felten and Appel:

As you have likely read in the news media, certain New Jersey election officials have stated that they plan to send to you one or more Sequoia Advantage voting machines for analysis. I want to make you aware that if the County does so, it violates their established Sequoia licensing Agreement for use of the voting system. Sequoia has also retained counsel to stop any infringement of our intellectual properties, including any non-compliant analysis. We will also take appropriate steps to protect against any publication of Sequoia software, its behavior, reports regarding same or any other infringement of our intellectual property.

Very truly yours,
Edwin Smith
VP, Compliance/Quality/Certification
Sequoia Voting Systems

[contact information and boilerplate redacted]

Comments

  1. sir_flexalot says

    the only agreements i’m aware of that prevent testing that way are religious, i.e. “you can’t test God”. How that would apply to a public vote seems so foreign to the purpose of this country, the gas-powered cranks on the graves of our founding fathers may still not roll them over fast enough.

  2. JohnLopresti says

    I appreciate the span of the comments. Seeing the problem originated with a progressive vote oversight official’s problemsolving efforts in the HAVA age in which we have embarked since 2002, in Union County NJ, I felt a moment of reassurance at having made the decision to emigrate to CA from Plainfield, NJ some years back. A glimpse at the NJIT website, as well as recollection of some of the political hassles through which Tova Wang had to slog, reminded me of the complexities which persist in the interfaces between politics in the eastern seaboard states as national politics oscillate between the various solutions our political parties afford to us.

    HAVA was a good idea, but implementation has been marginally successful, at best, and then only when valuable, insightful, goal-oriented people hold statewide office. In CA’s last statewide election we were fortunate to enjoy an excellent outcome with respect to the state elections oversight official, who had years experience as a state senator overseeing electoral matters in her committee. Debra Bowen has worked wonders since her election, and in a difficult environment with a script reading governor spouting partisan slogans. Of couse, the morass of politics is far from the tinkerer’s wish to see what the sourcecode is doing, but politics are a worthwhile portal to creation of an atmosphere in which good, thorough work may be accomplished.

    It seems to me, the approach some expert and involved NJ people have taken, viewed in aggregate, is appropriate; the HAVA process is going to upset cronyist politics and business status quo, but law is a living ‘cloud’, so, to my worldview, there remains reason to hope, and to work, as the blog owner is doing, to optimize the chances the sunlight of open evaluation of vote technologies yet may shine in NJ and nationwide. Without characterizing the political assets or drawbacks of specific regions in the US too explicitly, I would expect NJ to resolve this impasse, and several machine vendors to spinoff new divisions, redesign their concepts from ground up, and help lead us and their own profit margins into better times; and that this will occur with positive benefits in NJ long before some of the backwater areas in the US which are locked in chains of backroom politics. NJ is a bit cautious, but that can have its merits for essential technology evaluation such as voting machine design.

    I suppose all of this is little solace as our election year nears its culm. But I think the colloquia which ensue, even throughout the 111th congress, will help NJ and many other states thru these rough spots. Someday, like DOCSIS, Frame Relay, Asynchronous Transmission Mode, and many other standards, voting machine technology will become a hybrid of sharing and proprietary approaches to code design, as well as chip fab, so our votes will count, and runtime will be banned. That’s my take; the problem is in the canvassing as much as in the hackability of the voting station.

  3. I am organizing a panel that would speak to the issue of security with the Sequoia Voting System. This would be held in Chicago in Sept. 2008.
    Anyone interested in being a part of the panel should contact me
    at .

  4. I could probably produce a workable system in less than a month using an 8052, a couple LCD/controller modules, an SRAM chip, a couple of edge connectors, and some discrete logic and miscellaneous electronics (power supply, buttons, speaker, headphone jack, etc.) Someone else would have to design the case and mechanics, since that’s outside my field of expertise, but I don’t see anything too difficult there.

    I would probably write the code in C, but I would avoid any constructs which would generate machine code that was hard to read. A reasonably-competent person should be able to confirm the correctness of the design in less than a month.

    The design would rely upon the following security assumptions:

    -1- At least one person supervising elections would be alert and honest, and would have sufficient ability to raise alarm, should they find foul play, as to present an effective deterrent.

    -2- The people supervising the elections would be able to physically seal things in such a way that the things could not be altered without them knowing about it.

    To be sure, there are places where those assumptions don’t hold, and where the system I would design would thus not be secure. On the other hand, I would suggest that when and to the extent that those assumptions don’t hold, no system is secure.

  5. Haven’t I seen this already? Oh right… Man of the Year! Well, it’s too late for Bush to step down now though 🙂 You have my support, Professor.

  6. This is an outrage, this “Edwin” should be arrested for treason for attempting to use intimidation to interfere with the free elections of the United States, and possibly [redacted content that violates the site’s standards for civil debate — Ed]

  7. Anonymous says

    it’s not fraud. it is beacuse the jackass developer who wrote that code did not realize that arrays indexes start at offset 0 and not 1. Most likely they have an array Votes[i] and when they go to tally they use what ever languages Votes.Length() function which returns the length of the array but the array is Zero based like:

    Vote[0] = Obama;
    Vote[1] = Obama;
    Vote[2] = Obama;
    Vote[3] = Obama;
    Vote[4] = Obama;

    Vote.Length(); will return 4 even though there are 5 votes.

  8. Hey Ed, could you get a copy of the licensing agreement and post it on line? I bet it has a lot of dubious clauses in it.

  9. Joshua Thompson says

    I think any voting software should be required by LAW to be open sourced, to allow multiple independent analysis of the code. Nobody should ever be allowed to deploy a voting machine built on precompiled binaries. Period. There is simply too much at stake here to trust this to a closed proprietary solution, especially when there are literally hordes of programmers out there that will write this open software for you, absolutely free of charge.

    Ideally this code would be modular, with a good test suite. The back end vote tallying code should be separate from the user interface; this allows both to be tested independently, and it allows customization of the UI components while minimizing the risk of messing up the core of the system. The voting software would be a complete installable package with a stripped down OS, based on an OSS kernel such as Linux or BSD. Make it run off of a read-only boot drive such as a CD or DVD, and just have a small writable flash volume for storing the actual data.

    Next, take some of the money that would have been wasted on these proprietary machines and put some IT folks on staff whose job it is to take one release of the code, certify it, compile it and deploy it before the election..

    The hard part is the hardware. You can’t use proprietary hardware or the whole point of having the OS and voting software open is lost. There is room here for private companies too, as integration and testing labs who can help with designing and certifying hardware/software combinations. The difference is that instead of then selling you these boxes, they’ll give you the list of parts to buy and the assembly instructions; the individual states could then source these parts themselves from multiple vendors, assemble them, and install their certified software build.

    I suspect part of the reason for the legal threats is not just to avoid divulging bugs but also to avoid having their paying customers finding out just how simple these things really are (especially compared to the price tag.) I wouldn’t be surprised if, like another famous line of voting machines, are based on something horrible like MS Access databases.

  10. I don’t know if any of you are programmers but I am sure that you can imagine that code that captures and sums how many times a particular button was pushed could be written by the average 9 year old, the only secret they could possibly have would be something far more complicated, something that would nudge votes in a certain direction at a below the radar level per machine, getting it to say that 1+1+1 = 3 is childs play but getting it to say it = 3.000002 is far more sophisticated. its not ‘dumb’ or just ‘oopsy’ bad math, computers dont do that, they do exactly what they are told. Like me you may have found that some very widely used software out there actually takes the users input to round 4.5 and comes up with 4, every time and question my assertion, but even though you asked for ’round’ the computer was using a different set of rules than you and did exactly what it was told.

    so these machines were told by smart people to lie, in favor of republicans and just to be safe, lets make it a tiny white lie.

    velkommen til amedika!

  11. I’ve yet to understand what makes these voting machines so difficult to write for. As a developer, I’ve thought up scenarios, and most (if not all of them) are simple to write.

    This is simple arithmatic, and tracking. To me, the issues arise in authenticating whether the person is a registered voter, and tracking votes. A responsible company would track the data and do nothing with it., yet voter security requires that the voter remains anonymous. So you’ll have all these voters going to a voting station (School / Church / whatever), presenting themselves to a person, who points them at a machine and authenticates the machine to have one vote (Machine generates some ID which is unique to that voter, but not tied to their identity in any way).

    This leads to a point of failure – human interaction is required, and people are flawed. We need to be trusted, yet our society trusts no one. We need to track whether the vote count is legitimate, yet we can’t track the people and tie it to their votes.

    And yet they fail on a regular basis. Can’t add seems to be the main issue. How does this happen? One person, one vote, add 1 to a “voted on” Category, and 1 to the “Total votes” item. Sum the categories, verify it matches the total votes. Total time to do so would be a tenth of a second. Get one to work perfectly. Repeat for 5 billion people.

  12. That email isn’t even from a lawyer. Shows how smart that VP is…

  13. In Australia, we use paper votes. Everyone has to vote.

    We write “1” “2” “3” and so on until we’ve filled it in for the lower house, and write a single number above the line to vote along party lines or write up to 1..100 (or more depending on the year) to vote independently in the senate.

    There are several Electoral Commissions, the federal Australian Electoral Commission, and various state electoral commissions, and these are fiercely independent of the politicians. They are charged with drawing up electoral boundaries based upon number of voters in a particular area with minimal boundary distortion to avoid gerrymandering. Everyone voting in the same election is counted using the same method so the principle of one person, one vote, one value is upheld. Compare with the 2000 US presidential election with 50 different voting mechanisms.

    The majority of votes are now machine read using character recognition. Any non-machine readable are hand counted and added to the total. The entire nation votes during a single election day (usually a Saturday). We almost always know who has won around three hours after the election booths close.

    If there is a dispute with a result, the entire result is hand counted by independent Electoral Commission workers, watched by representatives of interested parties. This usually sorts out those results in a day or two, and they rarely affect the outcome of the election. Only a tiny fraction are re-counted a third time once all pre-polling votes are in.

    Why you folks use voting machines is beyond me. Expensive, unreliable, and unverifiable. Why those voting machine companies are allowed to contribute to political parties is beyond reason – it’s corruption waiting to happen.

    The voting process has to be transparent, verifiable at all stages, absolutely adhere to one person – one vote – one value, and be robust in the face of significant well known voting attacks. These machines fail at every single element of this most basic of tests.

    Andrew

  14. Thank you for fighting for our democratic process.

  15. Would you need more information (I am not working for Sequoia, nor voting). Each company at whatever level of implication (army, vote, government…) has the right to protect its Intellectual Property. Even just for protection against other competitor companies. That being said, there is a number of ways to make sure there is no fraud. Many of them are listed in a press release (on sequoia web site). I personally did not checked if they really did what they said they did, but if so, there is nothing to argue about.

    http://www.sequoiavote.com/press.php?ID=51

  16. Since when are the instruments used at the very nexus of American democracy subject to the whimsical notions of intellectual property and contract law? Let the machine be picked apart to the people’s content.

  17. Anonymous says

    Zero based array arithmatic. They need to hire some testers

  18. Glen Turner says

    Ben from Canada asks “What the F$%# is wrong with the world biggest democracy?”

    The world’s most populous democracy is India, not the USA.

  19. Does breach of contract to reveal prior breach of contract still constitute breach of contract?

    It was the foolish negotiators in the states and counties who executed these contracts without listening to the activists telling them otherwise – it is us, our government, who made the raw deal.

    If we thought the company would just pull its machines and go away, that would be great. But our $3 billion would not be enough – they’d all be willing to sue for breach, I think.

  20. Why would anyone want to buy or rent something that they don’t know how it works? Fire ’em and hire someone that will let you look at the software and machine inside out and outside in.

  21. Why “machines” at all? The first time I voted it was Check-the-box you want to vote for, tear off the stub, put your ballot in one box and the stub in the other and the ballots all got counted in public five minutes after closing time. I’ve lived all over this country and I don’t think any voting district is so large as to preclude a hand count of the ballots once every two years. No chads, no optical scans, nothing like that, just a simple two part ballot and a #2 pencil. Bring back democracy, get rid of big business!

  22. R. S. Buchanan says

    I’m confused: how can any license be legally valid that purports to void any right on the part of the licensee to test the product for fitness to purpose? How can they claim any sort of warranty of merchantability in such a situation? I suspect the answer is that they cannot, and NJ election officials need to grow a pair.

  23. I just read this via CNET. Forget them, tinker away. I wanted you to know I think an independent assessment is not on;y a good idea, it’s a MUST. Carry on!

    The reality is that the electronic voting machines everyday seem to be more and more staged and rigged based on independent and creditable news.

  24. The Sage of Brigadoon says

    ‘Sequoia will have positively NO non-compliant analysis…!!!’

  25. If “a process of voting” which is merely a process of adding numbers and storing them is ‘proprietary’, then there is a problem–yes, I realize this is not a patent, but a claim for copyright, but copyright doesn’t cover small works, and copyright doesn’t give them right to (try) to prevent people from looking at the code–that is the contract. But the contract was mis-written (i.e. not enforcable), or is being misunderstood by one or both parties.

    …Or there is some mass conspiracy that the heads of this analyiss aren’t aware of yet.

  26. What a funny letter.

    First, I don’t see why Sequoia would think that the recipients of that letter would care whether the action would violate NJ’s agreement with Sequoia.

    Second, the recipients are people who are well-documented to be aware of the Fair Use aspect of what they do. I’m uncertain why anyone at Sequoia would think there is the slightest chance of intimidating them in that manner.

    Third, if the result of the testing is unfavorable, then New Jersey would probably have plenty of basis to bring a voter fraud criminal action against Sequoia and everyone involved in the defects in question (from the C-levels down to the programmers), and the results would not be published until Sequoia as an entity ceased to exist. Thus, null threat.

    Fourth, I’m now *really* glad I didn’t take up that offer to work for Sequoia as a software developer full-time instead of finishing college. While I wouldn’t have participated in any voter fraud efforts, avoiding that would be more stress than I’d want. I’ve never had to change jobs to avoid criminal liability, and that makes me happy.

  27. I’m a computer engineer by trade. If I was given a contract to program voting machines, it would be very simple matter to:

    a) print a receipt matching your vote, but record something else on the machine’s hard drive
    b) record all votes correctly, but add them up and report or forward them in a conveniently wrong way

    I could also publish source code that looked legitimate and say “here it is, everyone check it”, but then run different code during the election. Even if the source is published and experts bless it, how can a voter or election official know that this blessed code is running on the machine during their vote?

    I think we should go back to real people counting paper ballots independently. It’s very easy to coerce machines to do something unethical (just look at all the computer viruses installed on the typical PC without their owners’ consent). But it seems it would be harder to coerce a group of (independent) people to do a coordinated unethical thing. Even if the people were only semi-ethical individually, coordinating them for a specific outcome would seem hard to me.

  28. I feel sorry for all residents in this state. Is this a case of the vendor being scared of bugs coming to light or is it more sinister with them worried about a back door being found? I guess NJ voters will never get to know if their votes are actually counted properly. Is this the type of democracy that young men from NJ are dying for in far away lands?

  29. Doug Taylor says

    I guess a lawyer can put pretty much any set of ridiculous demands into an email and sending it off to pretty much anyone on the net. Whether the demands are enforceable is ultimately up to the courts. This one looks like a fairly typical “My bosses don’t like you so they asked me to write a nasty letter” attempt at bullying. I’d say, print it out and run it through a shredder for entertainment.

  30. Ben from Canada says

    What the F$%# is wrong with the world biggest democracy?

    For gods sake, you people run around the world like it’s a crusade spreading the idea of democracy.

    But you can’t even prove that the electoral process in the grand old USA isn’t corrupted.

    Someone on Capitol Hill with a real set of balls, should demand that all electronic voting machines be inspected by independent 3rd party (ie. someone like Edward Felten), and grant full immunity from frivolous lawsuits and threat of lawsuits to do so.

    It’s simply embarassing reading about the shenanigans the voting machine companies and their politician friends are going through to control the outcome of said democracy.

  31. i am shocked; shocked to find that cheating is going on in here!

  32. How about a law to insure that all voting machines are subjected to both independent and public testing? It should help to deal with unscrupulous practices like the one above and it can help to reassure voters, even if there is nothing wrong.

  33. Richard Chapman says

    Somehow, in the last 200 or so years, “We the people” became “We the corporations”. A voting machine is the pivot point of Democracy. If it becomes owned, we become owned. But no matter what digital device we use, if the code that operates it is private, then everything we do with that device is subject to the whims and fancies of the true owner.

  34. Bouwer e-voting-machine bedreigt onderzoeker…

    In de voortgaande strijd voor open democratische verkiezingen nu het nogal ‘enge’ bericht over het Amerikaanse Sequoia, bouwer van elektronische stemmachines, die een dreigende email stuurt naar Ed Felten, professor aan de universiteit van …

  35. Weren’t there Sequoia’s voting machines in Florida when G. Bush won there by several hundreds votes or so in the state were his brother was a governor??? If “yes”, then Hi you Democrates!!! it is time to ban using non-OS software or hardware equipment for voting, otherwise US becomes worse democracy then Iran or North Korea not speaking about Russia!!! Do you want to lose having 3/4 of total US votes??? If you do NOT want, then ban non-OS for governmental and evidential use ASAP as Brazil has already done it!!!

  36. This is why my county went to vote-by-mail for all elections. They did an analysis and found that no electronic voting system was reliable; secure; or cost-effective.

  37. Take a look at Sequoia’s legal to see the bunch of bullshit going on. Alan Cohen, you are way off on this. They have the most unenforceable legal disclaimer I have ever seen…..additionally who has a disclaimer this huge and unenforceable for a WEBSITE?

    http://www.sequoiavote.com/legal.html

    Nobody ever agrees to a dispute resolution merely by accessing a website. (proven in court to not be able to be forced on people) Nobody ever agrees to have terms of use changed without notification (legal requirement there).

  38. Alan B. Cohen says

    Correction to my post above;
    I would expect they would want to see/comment on your results before you publish, but that is obviously up to any agreement you make with them. Another thought, they may also accept a confidentiality agreement between you and the Sec. of State of NJ that protects the confidentiality of their code.

  39. Alan B. Cohen says

    Seems to me to be a simple boilerplate stating that Sequoia wants its IP rights protected. Why not contact them to see if all they want is a license with YOU to review and let them examine your results. If you find a problem AND they are willing to fix it, everybody benefits.

  40. “it violates their established Sequoia licensing Agreement for use of the voting system.”

    If that statement is true then the Sec of State (or whoever the NJ voting authority is) should be held accountable for signing off on a licensing agreement that does not allow for independent analysis and verification of a system that is paid for by tax dollars, used for an essential public purpose, and MUST have the public trust.

  41. Do be aware, folks, that half of this is the blow back of technological ignorance. The “software intellectual property” protections they are citing was purportedly put into law as a way to stop hackers from cracking DVDs. Unfortunately, it also meant that most of us who use linux computers instead of Windows or Mac are committing a felony punishable by $100,000 fine and five years in prison if we PLAY a DVD on our computer that we BOUGHT. When I complained about this to my senator Saint Paul Wellstone at the time he wrote back that it was a great law and he’d sign it again if he had it to do over.

  42. As a voter, I have complete confidence that Prof. Felten’s analysis will be complete, correct and very interesting. As to the objections of Sequoia, I say: “tough!”.

    Come on, guys, it’s a voting machine. There is very little “intellectual property” involved in building an embedded system to put up a display, register and record touchscreen hits and save it to a memory card. It’s all clearly outweighed by the right of the people to have some confidence that their votes are being correctly registered in a way resistant to tampering.

    That said, I’m thoroughly against any voting system that doesn’t involve physical, human countable ballots. There’s just too much room for hanky-panky. Voting is a clear example of a process where the KISS principle should apply, both for transparency and accuracy.

    Touchscreen voting machines are a “solution” to a problem that never existed. They are highly profitable for Diebold and Sequoia and their ilk, but the design is too complex for the task, making them unreliable. Would you put a laptop in a warehouse for a year (or two) and then expect it to work perfectly for an entire day when you turned it back on?

  43. New Jersey election officials have stated that they plan to send to you one or more Sequoia Advantage voting machines for analysis. I want to make you aware that if the County does so, it violates their established Sequoia licensing Agreement for use of the voting system.

    This simple demand fully explains why we are not yet ready for electronic voting. The fact that government officials were given the authority to agree to such a crippling license (thus putting the entire public at risk), spells out that the voting process must be handled with extreme resistance to new technology.

  44. Then they could always buy some of those Sequoia offered to buy back from Florida for $1 each to allow examination by Ed Felton. They demonstrated the value is negligible.

    New Jersey currently has a State Voting Machine Examination Committee and reserves the right to require additional documentation from the vendor as well as State testing of any voting system as part of the examination and certification process. There is no guarantee this was the case when they obtained the Sequoia Advantage touch screens. You’d think that 17USC § 107. Limitations on exclusive rights: Fair use, and the doctrine of first sale would limit copyright law based restrictions, implying a contractual issue of some sort, or a bluff.

    You could also note that New Jersey is currently soliciting applications to certify centralized high speed optical scanning voting systems.

    “Pursuant to N.J.S.A. 19:48-2 and N.J.S.A. 19:53A-4, Election Laws of New Jersey, no voting system can be used for any election in the State unless it has been certified by the Attorney General, the Chief Election Official of New Jersey. The purpose of the certification process is to assure that a voting system meets all federal and state law requirements.
    http://www.state.nj.us/lps/elections/voter_issues/Notice-of-Invitation-to-Apply-1.8.08.pdf

    Should Sequoia choose not to allow certification those 11,081 touch screen voting machines may just be replaced with paper ballots, as not meeting state and federal law requirements. It doesn’t seem likely Sequoia’s refusal to allow certification through whatever process the State of New Jersey deems would likely succeed in a court of law. They should think of it as another chance – the State of New Jersey could adopt the California’s study, after peer review of the methods.

  45. Well, if they want to play it that way:

    Sender:
    To: Smith, Ed [address redacted]@sequoiavote.com
    CC:
    Subject: Re: Sequoia Advantage voting machines from New Jersey

    Dear Mr. Smith,

    Very well. If that is the official opinion of Sequoia Voting Machines, then our professional opinion to the election officials in New Jersey will be that Sequoia Advantage voting machines are not suitable for use in elections in the state of New Jersey as they cannot be independently validated and should be returned to Sequoia Voting Machines for a full refund.

    Have a nice day.

  46. In the trenches says

    If anyone actually read the CA source code review, they know why Sequoia would send such a letter.

    If anyone actually worked with the Sequoia system, they wonder why the ill-guided letter from Sequoia was not stronger!

  47. Ultimately our votes are being counted in SECRET. No one I know is comfortable with that. Also, since when were our elections usurped by vendors and HAVA lobbyists? I thought elections belonged to the People. I did not know that private corporations now run our elections FOR us: I guess they think we are so stupid we are incapable of counting OUR OWN PAPER BALLOTS. It is time for Electronic voting to be a thing of the past. Send all the machines on a one way ticket to the LANDFILL.

  48. Remember that John Gorka song, “I’m from New Jersey”? There’s a line, “It’s like Ohio, but even more so, Imagine that…”

  49. From my perspective, Sequoia is teetering on the edge of bankruptcy and certainly may not withstand the financial downturn that is 2009. Sequoia is misguided in manifesting a confrontational business posture with most clients, in most states, by their senior management or those individuals who pretend to be senior management.

    Is Mr. Smith Senior management? I think not – Senior management would not say such a thing. Does Mr. Smith have 5 years of election experience, I think not or he would not say such a thing.

    A number of posts on this site indicate that the obstacles incumbent upon one who wants to peruse an action against the State, or it’s agents, are insurmountable. They are – been there, done that, and spent a fortune doing so (and won). That said, how many individuals who have read all this have or are considered filing a FOI request for this information for the contract of which Mr. Smith Speaks? I checked, he (Mr. Smith) is registered in no state as an attorney (and that was a bitch in and of itself).

    Has anyone reviewed the Sequoia web site? Do they not represent that they will help in the testing of their system? Will the web site be changed by 8:00 tomorrow Eastern time?

    I worked on the system in question (it has big and real problems), went to a good school (ranked better then Princeton), have 20 years of experience in elections (in a real environment), and have to question why we, as a community, cannot solve this simple issue?

    What does Democracy mean anyway, “mob rule”? Are we not talking about a swing state?

  50. So let me understand it – for the computer phobic public:

    What the legislators want us to believe:

    we have a solid/transparent system that ensures a valid count and security of voting

    Reality:

    This company is scarred sh*t less of an academic analysis of their system

  51. @catalin: “paper trail” does not necessarily entail a receipt that you carry off site. What it could mean is this:
    1) every voter gets an anonymous printout of their vote, so they can check that their vote has been recorded according to their wish.
    2) every printout goes into a ballot box – still anonymously
    3) after closing, the voting machine tallies the votes, more quickly than hand counts – that is the one thing voting machines are good for
    4) after the election, the results of randomly selected (AFTER the vote) voting districts are compared to the printouts collected, to see if any manipulation happened

  52. Why is this company asserting that:

    1. Professors Felten and Appel are bound by the “Sequoia licensing Agreement” between Sequoia and New Jersey? They are not. Its not possible to violate an agreement that does not exist.

    2. They have a right to prevent “non-compliant analysis”. Compliant with what, an agreement that Professors Felten and Appel have never been party to and are not bound by? Ridiculous.

    Why are so many shady and seemingly criminal acts being perpetrated around voting machines that are going uninvestigated and unprosecuted? Not only Diebold ( or whatever new pseudonym they call themselves these days to try and conceal their identities.. sure seems like something a legit company would do! ), but now Sequoia as well.

  53. Just send the machines to Florida, the votes don’t count here anyway. Remember hanging chads and early primaries void our votes anyway.

  54. “I don’t think this has anything to do with “intellectual property” and rather with contract law. By default, when you buy software you own your copy. This means you can run it however you want, as well as to read your copy of the software (for example via a disassembler).”

    Hi Lior,

    This is not strictly true, for example if you disassemble code and print out the disassembly for analysis.

    A commonly cited test case is Sega Vs. Accolade, in which Accolade reverse-engineered Sega’s code to figure out how to interface with a Sega Genesis.

    Accolade’s defense argued that (1) the object code was not really copyrightable, (2) transient copies like printouts did not fall under copyright (they weren’t publishing or distributing them or such); and (3) it’s all fair use anyway. The judge did not buy (1) or (2), concluding that transient copies of disassembled code were subject to copyright restriction. However, he concluded that it clearly fell under fair use.

    I think if Sequoia believes copyright restricts analysis of their software, they are uninformed and don’t have a case; on the other hand, it doesn’t hurt to consult with a lawyer and make sure to reverse-engineer with care.

  55. If this is really just a contract matter… why can’t NJ just breach the contract?

    I seem to remember that there is a breach of contract defense available when the contract violates public policy. Does NJ recognize this defense? Do the terms of this license agreement rise to the level of a public policy breach as recognized in the courts?

  56. Gunslinger says

    Shhhhhh! All this independant thought could rouse the anger of our shadowy overlords whom are wiser than We The People.
    Best to just go back to watching American Idol and tabloid news.

  57. Send me one.

  58. joe roberts says

    If a company wishes to create a system for any mass public use/good then if the taxpayers or those who represent them wish to verify/validate any and all claims by said company about their good/service the company should be fully willing to do so. If they are not then the contract with said entity should be considered Null and Void under the best interest of the PEOPLE. As for IP claims about 3rd party testing this si not common go home and dig out your EULA for any and all Microsoft products, they expressly forbid testing without their consent. Is it right? NO But that’s the kind of stuff that is allowed to happen when people care more about the almighty dollar than the good of the people.
    Remember for the Dollar and by the Dollar is how this country is run.

  59. THEY ARE AF5RAID OF YOU REVEALING THE TRUTH!

  60. Independent research is almost always covered under fair use, as far as I know.

  61. To hell with their license. Crack their code open and show the world all the bugs.

    If Sequoia didn’t have problems with their voting machines, they would let people inspect them.

  62. One thing for sure is that hackers never violate licensing agreements…

  63. Are they trying to suggest contractual interference? ie: you are enticing NJ to violate their contract?

  64. Brian Kemp says

    You could always just get hired as an Employee/Contractor by the State of New Jersey temporarily, which would mean that you’re no longer independent and instead working for the government, which would probably neatly work around the licensing agreement.

    Or the NJ government could just make independent authorization a criteria for bidding, in which case that would also neatly work around the problem. 🙂

  65. In fairness to Sequoia, I can understand them not wanting proprietary technological information reaching the public domain, and you could even argue that it’s a technical violation of the warranty. But they don’t really think you’re foolish enough to publish anything a rival firm could exploit, do they?

  66. It should be easy enough, by law or regulation, to make availability for testing by whoever the state chooses a prerequisite for continued certification. But it does surprise me a little that Sequoia was stupid enough to send a letter that would reflect so badly on them.

  67. Carlie Coats says

    IANAL, but…
    In the initial Supreme Court decision establishing the doctrine of Fair Use, the SC said that Congress is forbidden
    to establish a copyright regime so strict as to damage First Amendment Freedom of the Press and Freedom of Speech;
    published reviews containing quoted extracts were the specific matter in question.
    By 14th Amendment extension, the same applies to the States and their contract laws.
    IMNHO, Sequoia’s position is Constitutionally unsupportable.

  68. We went through much of the same issue as a prelude to the California Top-to-Bottom Review. There were never any explicit threatening letters like the one sent to Princeton, but we fully expected something like this could well have happened. Consequently, we were all employed by the University of California in order for the State of California to shield us all from any liability if any of the vendors might have decided to try suing us personally. A lot of lawyers spent a lot of time and energy concluding that this was the proper way to do the review.

    If/when a review occurs in New Jersey, particularly given this vendor’s striking lack of cooperation, I imagine similar advance legal preparations will need to take place.

  69. I can see where a company wants to protect it’s trade secrets from competitors. But voting machine technology must not be transparent to the public if we are to protect the sanctity of democratic processes. A judge may not be able to see this correctly, but if you get enough politicians or government organizations behind you on this, you should be able to get Sequoia to at least back down.

  70. Out with the Diebold, in with the Diebnew…

  71. I wish you had left in the contact information.

  72. disgusting,i hope youre not going to listen to them. that would be a shame

  73. Too bizarre. I can’t wait until Ford sends a letter like this to Car & Driver, or Maytag sends one to Consumer’s Reports.

    Basically what they’re saying is “Our product is so bad that if you mention it aloud, we’ll SUE!!”

  74. What folamir describes would break receipt-freeness, one of the desirable security properties of any election system. With their receipts voters could accurately prove to third third-parties how they voted, and this would make them more susceptible to coercion. Not providing such a receipt for the “right” candidate would be seen as disobedience and could be punished by a coercer. Voters would be also able to sell their votes this way.

  75. Michael Seery says

    I vouch $100 for the legal defense fund, should one be formed.

  76. Wow. Our democracy is now the IP of BigCos……

  77. Pure scare tactics. If there is a valid legal issue, it’s with the state not with the researcher.

    The only way to make voting secure is open is to make it open. The system I’d like to see gives everyone who votes a receipt with a unique code number identifying their vote. The consumer should be able to go to a website after the election and check that their vote was actually counted. If votes go missing, people will find out.

  78. Although Sequoia’s behavior is legally excusable (?), it is a force against fair elections. This is not as much an issue of open-versus-closed-source systems, as an assertion of power by the e-voting industry over the state.

    I won’t make the leap that Sequoia’s reasoning is that the machine is a dump, though. However, I’ll remain skeptical of its performance unless trusted third parties give it their approval.

  79. I am William Bunker and I am running for House of Representatives in Massachusetts.

    Sequoia ‘Advantage’ voting machines and any electronic voting machine have a Constitutional liability to publish their source code and all software and to print out a paper trail when recording votes.

    Failing to do this is grounds for a junk election and interfering in this process is arguably vote fraud. Depending on the circumstances, legal action to attempt to conceal vote machine source code may itself be fraudulent in nature.

    We’ve all seen a wealth of evidence supporting probable and in some cases unmistakable intentional electronic voting fraud. Minimizing this risk is important, and prosecuting suspects of this is important. Due to the free and public nature of elections companies have no right to expect privacy of voting machine source code.

    See my page at http://www.myspace.com/peaceisthesymptom and watch the video of the programmer testifying before congress. He says that he was asked to produce a program to discreetly change vote totals in public elections. This kind of behavior by companies is unacceptable.

    I will introduce legislation to investigate and prosecute electronic voting companies and ask our judiciaries to find cases against them if elected.

  80. This looks rather sad for the voting machine company. “No! Don’t show it to Ed Felten, he’ll rip it apart and embarrass us!” Typically when experts review products, the makers usually say something like, “hey, here’s a free one, please test it out and let us know what you think.” Not sure why this is any different.

    Anyway there’s a great opportunity for Princeton Undergrads looking for a project in Prof Felten’s class (a class I was lucky enough to take while at Princeton). Design a secure voting machine that passes muster. You’ll have customers lined up, as they already recognize the Felten lab as an authority on this stuff. Perhaps it would be a semester-long project. Let me know if you actually do it! 😉

  81. The s*** is going to hit the fan. Probably 2 nights from now we’ll see this story on Nightline.

    Voters are really concerned about the integrity of the process, with so much at stake in November. Even my 70 year old mother (*especially* my 70 year old mother) is concerned & interested in this topic.

    Sequoia — methinks thou doth protest too loud. (Not really used appropriately, but I do think they are protesting too loud (too loud = at all) )

  82. Another item for the Gallery Of Legal Threats 🙁

  83. Problem for Sequoia is that if their contract is with the State of New Jersey, contract law may do them no good. Unless a State lets you sue them, you can’t sue a State (with some exceptions stemming from the 14th amendment).

    Of course, if they are trying to sue the state under intellectual property laws they may be out of luck as well. A State has immunity from patent and trademark suits filed by anyone but another state or the feds, presumably also from copyright laws.

    I don’t know whether than immunity extends to independent contractors hired by the state though.

  84. Clever wording, by the way: “publication of Sequoia software” is, indeed, copyright infringement, but there is no hint of any vote-machine reviewer having ever done that. “Publication of … its behaviour”, on the other hand, is perfectcly legal (perhaps Sequoia intends to claim that the bugs in the software constitute trade secrets?).

  85. I don’t think this has anything to do with “intellectual property” and rather with contract law. By default, when you buy software you own your copy. This means you can run it however you want, as well as to read your copy of the software (for example via a disassembler). You are free to accept more restrictive licensing terms when buying the software, but it’s your fault if you do. If, indeed, the contract says what the laywers say it does, then NJ has a serious problem.

  86. Sounds like a desperate attempt to stop an independent evaluation of their system, and this alone tells a lot about how (in)secure their system might be. I wonder whether they can really enforce their “intellectual property” (whatever they mean by that) without totally destroying their credibility.

  87. Why would the election officials agree to no independent testing given the recent voting machine news? Couldn’t they simply force the Sequoia and the like to allow this? What are the voting systems companies going to do… find customers outside the government?

  88. But, Michael, of *course* intellectual property is *much* more important than the integrity of the voting process!

    I’m not entirely sure why *anyone* trusts any Sequoia machines *anywhere* near any voting system, or indeed any system for which security is at all important. Emails like this make it plainer than ever that they’re part of the problem space, not the solution space…

  89. Michael Donnelly says

    Wow, I’m stunned. Somehow, this voting machine thing is all about… intellectual property? Could a company really try to use copyright law as a weapon when it clearly has no place? To quote:

    “infringement of our intellectual properties, including any non-compliant analysis”

    Oh, brother.

    (/sarcasm)

  90. Ben, it seems that the policy doesn’t prohibit testing … it prohibits independent testing. (Which is a worse prohibition. It means you can get biased testing results out there, but not unbiased results)

  91. I think a policy like that should make the use of said machine a no go in any election at any level.

  92. So will we get a separate post on a) how you have/will respond and b) the implications of a licensing agreement that prohibits testing.