Tangled Webs

Issue 9.1
Feb 24, 2004

Vote Early Vote Often

The last American presidential election resulted in both endless debates over "dangling chad" and a renewed interest in electronic voting. In theory, replacing punched paper with computerized counts should lead to more accurate and transparent elections, but things are not working out that way.

As America gears up for this year's presidential election, half a dozen companies are jockeying for position in the potently lucrative voting systems market, with Diebold one of the current leaders. Things do not look good for democracy so far.

Last January, students from Swarthmore College broke into Diebold's corporate servers, and posted election software and internal memos on the Internet. The material was startling. Not only did the election system contain sophomoric programming errors and lack basic security precautions, but Diebold executives had no intention of fixing these problems.

Diebold's response was to slap the students with a lawsuit for, of all things, copyright infringement. As 2003 progressed and the scandal grew, the incriminating documents were copied all over the Internet, and Diebold expanded its legal harassment to dozens of individuals and ISPs. Public outrage and political pressure eventually forced Diebold to drop the lawsuits and promise to fix the election software.

Further doubt was cast in December when California discovered that Diebold had installed uncertified software in every single county that used their voting machines. A similar breach had occurred in the Georgia's mid-term elections a year earlier. In both cases, the uncertified software was almost certainly bug fixes and upgrades to the certified version. However, the fact that the software had defects severe enough to require last-minute (and technically illegal) patches demonstrates that it was in no condition to be used. Furthermore, that the uncertified software was not discovered until well after the elections indicates a deeply flawed oversight procedure.

In my opinion, however, the last straw occurred this January. The State of Maryland hired independent security experts Raba Technologies to evaluate the Diebold system for use in their March primaries. The results were typical of all such evaluations to date. Researchers were able to vote multiple times, take administrative control of the machines and change votes that had already been cast. William Arbaugh, who worked on the study commented, "I was really surprised with the totality of the problems we found. Just about everywhere we looked we found them."

Clearly the system is not yet ready. The responsible thing to do would be to use the old voting systems in March and evaluate Diebold's again for a future election. Surprisingly, however, Karl Aro, director of Maryland's legislative services department, said of the scathing report "It is a validation that the system is ready to work in March." And Diebold issued a press release claiming that the report "confirmed the accuracy and security" of their systems.

Those Who Count The Vote Rule

So what's going on here? A few Democrats think the answer lies in Diebold's secrecy and nervously point to the company's strong ties to the Republican Party and the outspoken political views of its board members.

While election fraud is not exactly unheard of in American politics, I think Diebold is acting as any corporation would in this case. They are doing their best to maximize shareholder value. Going the extra mile to make a secure and robust election system is expensive and would not earn them any more money. Likewise, Maryland bureaucrats have spent millions in taxpayer funds, and some probably have their careers on the line. There is nothing nefarious. It's simply a situation in which both parties must be able to declare the results a success regardless of the actual outcome.

And therein lies the problem. Fair and open elections are the foundation of democracy. They are far more important than a career bureaucrat getting his next promotion or a corporation hitting its next earnings target. It seems that America's capitalist and democratic ideals are fundamentally divergent here, and we have been going down the wrong path.

Fortunately, the solution is simple, inexpensive and, well, democratic. First, all software developed for or used in election systems should be either placed in the public domain or made available under a public license. Anyone who so desires should be able to inspect the software, make suggestions for improvement, or even donate their time to enhance it.

The hardware required to run election software would have to be generic enough so that it could be procured from multiple suppliers. In fact, to further reduce the chance of fraud, multiple hardware suppliers should probably be required.

The government's job in this process would be to set the requirements, to partially fund development, and to make the final decision on which software gets certified. The government certification board would sign certified software with a digital key to ensure that only unmodified certified software could be installed on voting machines. Most important, every step of the process, from every line of source code submitted to the results of every review, would be disclosed to the public. There is no need for secrecy over the democratic process.

The final step comes after the polls close. Copies of the data files from all electronic voting machines would be collected and placed on the Internet for everyone to see and inspect. With a properly designed system this would make vote tampering all but impossible and eliminate the need for anyone to call for a recount since votes could be counted by anyone who desired to do so.

All In Favor, Say "Aye"

Many people seem somewhat uncomfortable with this degree of transparency, and in some lights it does seem extreme. However, one of the fundamental principles of democracy is that the citizenry, not the government, is responsible for safeguarding the democratic process, and this system would give them the tools to do that.

Open election software could be developed at higher quality and a fraction of the cost of its proprietary counterparts. These systems are not particularly complex, and the success of Linux, Apache and hundreds of other open source software projects have shown that many of the world's best programmers willingly donate their time to projects they consider worthwhile. The chance to help create the mechanism that underlies modern democracy would have an irresistible appeal to many.

Since the software would be free and the required hardware inexpensive, it could be used by everyone from schools electing a class president to budding democracies holding their first open elections. And things would not stop there.

If America can develop a truly open and transparent electronic voting system, it would likely eventually be adapted and used in democratic nations all over the world. The job of election monitors would become easier since they would be assisted by millions of eyes poring over the results.

Of course, like modern democracy itself, as the system is adopted, changed and occasionally improved upon it would cease to be an American system. Like democracy itself, it would be owned by the entire world.

[ Home Page] [ Back to Index ] [ Previous Issue ] [ Next Issue ]

© Copyright 2004, Tim Romero, t3@t3.org
This article first appeared in the February 18, 2004 edition of The Japan Times.
Tangled Webs may be distributed freely provided this copyright notice is included.
The Tangled Webs Archive is located at http://www.t3.org/tangledwebs