By Teresa Hommel, 2/9/05



Attorney Generalís report

misunderstands computers and evote technology




Eliot Spitzerís report ďVoting Matters II: No Time to Waste,Ē published February 7, 2005, is disappointing. The New York Attorney General deals well with the familiar issues of voter access to the ballot, but misunderstands the new issue -- problems with evoting -- and supports ineffective solutions.


Elections are about votes, not computers.


Election integrity requires vote recording and vote counting to be meaningfully observable by non-technical, multipartisan observers.


Reading software (or hiring a computer scientist to do it), as suggested on page 15 of Spitzerís report, is ineffective to evaluate election integrity. First, no one can guarantee that the software to be read is the same as the software in the evote machines on election day. Second, no computer scientists will guarantee that reading software enables them to find all errors or fraud.


Current bills in Albany require escrow of software, but that is a slight-of-hand game. No one knows if the escrowed software was used during the election. Because data files and other software in the computer can alter how the election software works, as well as electronically-recorded ballots and tallies, the examination of escrowed software cannot safeguard an election.


A random computer test on election day, as suggested on page 15 of Spitzerís report, is also ineffective. Who performs the test? What non-technical election observer would understand the procedure? What exactly is being tested? (It is a student-level exercise to determine whether software is being tested or used for real interactive work, and any system can alter itself during and after the test.) To better understand this point, please run some tests and elections with the Fraudulent Voting Machine at Its machine tests always work, but its real elections are fixed.


Once a lever machine is set up for an election, it cannot alter itself. Once a paper ballot is marked and cast into a ballot box, it cannot change itself. But computers can alter their own software. A data file can act as software and alter the election software, recorded ballots, and tallies. Software in the same computer but outside of the election software can alter the election software, recorded ballots, and tallies.


Spitzerís report asserts that votes must be counted in a fair and uniform way. That standard isn't met by a surprise random recount of 3% of the votes and a "trust-the-computer" count of 97%, along with no legal requirement for accuracy, as current bills in Albany propose.


The tight deadlines of HAVA have prevented careful consideration of the nature and risks of computerized voting, as well as the threat to democracy when vote recording and vote counting are concealed as electronic procedures. Moreover, most lawyers, officials, and good government advocates are too busy to learn more about computers or keep up with the news of evote problems and failures. Spitzerís report, as well as the current bills in Albany A5 and S1809, are uninformed, as revealed by their escrow and audit provisions, as well as the Assembly bill's requirement that communications devices be "off" during election day in spite of the fact that few if any election personnel would be able to ascertain whether such a device is in the evote system, much less whether it is on or off, much less whether it has allowed alteration of software, ballots, or tallies in the few seconds before or after the election.


Computers can fraudulently provide any margin of victory and any randomized pattern of vote-switching. The recent Yale report showed that by switching merely one vote per evote machine, the winners in many races can be switched. Spitzerís report mentions tiny margins of votes (page 2) and close elections (page 22) but this concept is not meaningful when computers are used because the margin of victory can be determined before the election, and implemented afterward when the number of ballots cast is known. The concept of a close election is meaningful only with technologies where falsification of ballots requires huge coordinated efforts by many people in many election districts such as paper ballots or lever machines.


As long as so many officials misunderstand computers and evote technology, and Boards of Election do not have the expertise and resources to properly manage them, it is foolish to invest in such technology and turn over our democracy to it. Paper ballots marked by hand or ballot-marking machines for voters with disabilities are a superior election technology because it is less expensive and everyone understands how to safeguard the ballots and observe counting procedures.


Lever Machines


It is a political decision to fail to maintain our lever machines, and then demand that they be replaced because they are broken.


The lever machines were made to last 150 years with normal maintenance, and can be brought to nearly-new condition. New parts are available from International Election Systems Corp., 1550 Bridgeboro Road, PO Box 70, Edgewater Park NJ 08010.Telephone 609-871-2100.The president of the company is Richard Nowetner.


The constantly-repeated argument that the lever machines are "old" may appeal to young people, but in fact the lever machines are the easiest voting technology to safeguard, and the hardest to tamper with.


The HAVA requirement for a "permanent paper record with manual audit capacity" has been interpreted to be satisfied by an end-of-day paper printout generated by evote systems. Such a printout does not provide any security whatsoever for evote systems, because computers can produce authentic-looking printouts with entirely incorrect numbers.


Because a meaningless end-of-day printout from evote systems satisfies the HAVA audit requirement, there is no logical reason why this requirement should not be satisfied by the meaningful paper record that is manually created before multipartisan observers from lever machines at the end of an election day.


(Note that a meaningful audit of evote tallies would require a recount and discrepancy-reconciliation of voter-verified paper ballots. The need for a secret ballot prevents other methods of audits, such as tracking numbers.)


"Same Opportunity for Access and Participation"


Access and participation must not be limited to the voting experience. Focus on the voting experience in isolation, rather than on the entire election system and whether the votes are recorded and counted correctly, leads to placebo experience, unverified final tallies, and loss of voter confidence and election legitimacy.


The use of voter-verifiable paper audit trails solves one problem Ė the need to verify vote recording -- but does not solve the problem of observable vote counting procedures. Unless multipartisan observers can watch all votes being counted and all discrepancies between electronic and paper tallies being reconciled, the election provides only a voting experience. This has long been recognized: Boss Tweed said, "As long as I count the votes, what are you going to do about it?" Anastasio Samoza said, "You won the vote, but I won the count." Josef Stalin said, "It's not the vote that counts, it's who counts the votes."


Election integrity cannot be ensured by statewide similarity of voting technology, but by statewide requirements for meaningful observation by non-technical observers.


Privatization of elections


Spitzerís report says (page 9) that our electoral process including vote counting must be transparent and fair, and ensure the confidence of voters. Unless there is a routine 100% publicly-observable vote count, we are delegating to a corporation to perform a secret vote count, and putting the burden on candidates and parties to litigate for the right to obtain and observe a count of all the votes.


Spitzerís report recognizes the need for training of poll workers, but not the need for computer training for Board of Election personnel before using evote equipment. Lack of computer expertise would require defacto privatization, since all Boards of Election would have to use vendor technicians to maintain their equipment and run their election-day operations. This pattern is well-recognized in other states.


Elections are a function of the people


The public must have rights beyond access to a voting experience. When John Kerry refused to investigate or challenge what many voters believed were falsified results in November, 2004, the public had no standing to do anything except hold meetings and complain to the press, which didn't report the issues in detail.


Consistency of election integrity can be achieved by allowing or requiring multipartisan observation, not by turning over ballot-recording and vote-counting to an unobservable electronic process.




HAVA encouraged changes in election technology by offering money to the states, but did not mandate electronic voting.


HAVA does not require replacement of lever machines. Compliance can be achieved by giving each polling place a ballot-marking device accessible to voters with disabilities and non-English languages. The worthy goal of voter access to the ballot must not overshadow or replace the need for legitimacy of elections which, for evote and paper ballot systems, can only be achieved via public observation of the ballots once cast and 100% of vote-counting procedures.


A nationwide movement against evoting has arisen due to HAVA's inadequate standards for election integrity, the widespread practices of concealment of precinct tallies and all counting procedures in many states, and the refusal by local and state Boards of Election to acknowledge or deal with observed and documented failures of evote systems.


New York should learn from the problems in other states, and not switch to evoting when safer, less costly alternatives exist. We should not be gullible and accept federal money without fully knowledgeable, open consideration of whether the changes we make will improve or worsen our election system.