This provision clearly promotes the adoption of DREs. Many jurisdictions not using them in 2002 contemplated their purchase. HAVA authorized distribution of three billion dollars to the states, provided that the states submitted plans specifying use of the money consistent with identified HAVA requirements. An acceptable use was the replacement of punched card or mechanical lever voting machines with either optical-scan or DRE systems.
Following are some examples of concerns about integrity of DREs, beginning in 2003.
In early 2003, David L. Dill, professor of computer science at Stanford University, actively opposed, with public testimony, his local county's contemplated purchase of DREs. He began the Verified Voting Foundation and established a website at www.verifiedvoting.org. He posted on his website a resolution which was afterwards endorsed by many information systems professionals. The resolution includes the following:
"Computerized voting equipment is inherently subject to programming error, equipment malfunction and malicious tampering. It is therefore crucial that voting equipment provide a voter-verifiable audit trail, by which we mean a permanent record of each vote that can be checked for accuracy by the voter before the vote is submitted, and is difficult or impossible to alter after it has been checked ..."
In the summer of 2003, an experiment undertaken by Dr. Avi Rubin, a computer scientist at Johns Hopkins University, was widely reported by the media. Rubin and associates asserted, among other claims, that the "smartcard" given to each voter at polling stations in Maryland to be entered into a DRE machine in preparation for voting, could be easily duplicated; this would allow a voter to vote many times (Kohno et al., 2003). As a result, Maryland's government let two contracts to review security procedures, and some procedural changes were made. Rebuttal of the charges and corrections made were minimally publicized in the media.
In July, 2006, the possibility of "hacking" by just one person was highlighted in the presentation of a new report by the Brennan Center for Justice of New York University School of Law (Brennan Center, 2006). The limited conditions under which this "one person" might carry out the manipulation was not reported in the media. The report itself is considerably more detailed and provides an excellent description of necessary responses to vulnerabilities.
On July 31, 2006, a press release from the Open Voting Foundation of Granite Bay, California, about a particular DRE machine, was headlined: "Worst Ever Security Flaw Found in Diebold TS Voting Machine." The text stated:
"... it has been determined that with the flip of a single switch inside, the [Diebold TS] machine can behave in a completely different manner compared to the tested and certified version. ... According to [foundation president Alan] Dechert, 'If you have access to these machines and you want to rig an election, anything is possible ... and it could be done without leaving a trace. All you need is a screwdriver.' ..."
In December, 2005, the media reported an experiment by the Supervisor of Elections of Leon County, Florida. A computer security expert was permitted access to the internals of a computer used to count votes, and the expert altered the machine's program. Consequently, the machine reported an incorrect result in a mock election. The experiment was touted as another example of the vulnerability of vote-counting by computer, and it may be assumed that, as a result of the publicity, public confidence was further eroded.
What the media did not make clear in this case was that Leon County's voting system employs voter-filled-out optical-scan ballots. Thus, regardless of any malicious distortion of the computer program, the ballots themselves would be available to be recounted by hand or on an independently managed computer system to check the initially reported results. Limitations in computer security were highlighted but the available remedy of independent verification was ignored.
6. Testing and Assurance Activities of States Using DREs without Paper Trails
With opposition from professionals with the necessary technical credentials and the wide publicity given flaws in computer security, states using DREs at polling stations without paper trails have had to justify their procedures. Such states include Georgia, Louisiana, and Maryland.
In Georgia, Brit Williams, professor emeritus of computer sciences at Kennesaw State University (KSU) and technical expert for the state, has been actively defending his state's implementation. He spoke on the subject at a NIST technical symposium on December 10, 2003 and testified to the Committee on House Administration of the US House of Representatives on July 7, 2004, along with Kathy Rogers, Director of the Georgia Elections Division. He has noted that Georgia has created the Center for Elections Systems at KSU to provide support and independent testing for all of its 159 counties. Tens of thousands of voting system components are tested there and the staff continues to travel to each of Georgia's counties to independently test and validate all new equipment purchases. Georgia employs a method of testing its software to assure that no Trojan Horse program can switch votes. After an election and the closing of polls, the results for each machine are posted on the wall of the precinct. Any manipulation of the memory cards during the time they are transported to the central processing station would be found out. A more complete review of Williams' presentations is available (Saltman, 2006, pp. 206-207).
In Louisiana, protective measures take advantage of the fact that this state formerly employed lever machines statewide and had developed procedures to assure their correct operation. A recent communication from the office of the Secretary of State provided the following information:
"Every machine is tested prior to an election to assure it is working correctly and is voting correctly, i.e., it registers the vote to the correct candidate position and it produces a zero proof sheet. We have both a protective counter and public counter and each is checked on every machine before election. ... [The public counter] records the number of votes on that machine at that election. This can be checked with the precinct registers as our commissioners have voters sign the register before voting on the machine. ... A change in software cannot be made by [equipment vendor] Sequoia without our assistance/knowledge, nor could we make a change without Sequoia's assistance and knowledge. We have password-protected logons and we audit every logon and follow all activity in this regard. We also have limited access, of course, which depends on a person's level of authority in our department. ..."
Next Page 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9
(Note: You can view every article as one long page if you sign up as an Advocate Member, or higher).
