In December 2002 I received an email out of the blue, from David Dill. It said:
"I'm a computer science professor at Stanford. I'm aghast that places are actually using DRE machines with no paper trail. I would like to see if I can help to ensure that new voting machines are auditable."
And help he has. In the three years since that time, David Dill has started Verified Voting, testified before Congress and all across the country, and helped to establish ACCURATE, a multi-university, multi-million dollar National Science Foundation-funded project to improve voting equipment. Dave and I served together on the California Secretary of State's 2003 Touch Screen Voting Task Force, and I was honored to share the Electronic Frontier Foundation's 2004 Pioneer Award with him.
One of the many things I admire about Dave is his ability to understand human nature and politics. I also admire his commitment to working on voting technology problems, which are complicated and at times frustrating. But the years of toil have given David Dill a clear vision of where we need to go, as he explains in this piece, recently published on TomPaine.com and featured below.
Making Democracy Transparent
by David Dill
Public trust in our elections is eroding. While the general public still seems to accept election results, there is an undercurrent of bitterness that has grown tremendously over the last few years. There is a rapidly expanding body of literature on the Internet about the "stolen election of 2004," and several books on election fraud have recently been written. More are in the works.
Theories of widespread election fraud are highly debatable, to say the least. Some people enjoy that debate. I do not. It encourages a sense of hopelessness and consumes energy that could instead be focused on long-term changes that could give us elections we can trust.
The election fraud debate frames the problem incorrectly. The question should not be whether there is widespread election fraud. It should be: "Why should we trust the results of elections?" It's not good enough that election results be accurate. We have to know they are accurate - and we don't.
In a word, elections must be transparent. People must be able to assure themselves that the results are accurate through direct observation during the election and examination of evidence afterwards.
U.S. elections are far from transparent. Instead, winning candidates and election officials alike tend to put all their efforts into suppressing recounts. That attitude has led to increasing bitterness with each national election, at least since Florida 2000.
But we can conclusively win a debate about election transparency. And while making elections more transparent will be difficult, it is more feasible than solving many of our other national problems. All that is required for success is a long-term strategy and a commitment from many citizens at the grassroots level, since politicians and election officials are not going to solve the problems on their own.
Here are some initial thoughts on how we can do it. I propose a four part solution: We need to ensure that voting technology is transparent; election procedures need to be rethought to emphasize openness, security and checks and balances; election laws need to be revised to support these points and to make it easy for candidates to get reliable, manual recounts; finally, citizens need to participate in witnessing elections and making sure they are conducted properly.
Questions about voting technology have been in the spotlight in the last few years. The first concerns were about accuracy, inspired by the problems with punch cards in the 2000 election. The supposed solution to that problem lead to plans for the widespread adoption of paperless electronic voting. But paperless e-voting is totally opaque - no one can observe the handling of the (electronic) ballots. The hardware and software of modern computer systems are designed and built by thousands of specialists: Decades have passed since a single person could comprehend an entire computer system. As a result, there is no way to ensure that such voting systems are accurate or honest.
Right now, the only feasible solution to the insecurity of electronic voting is a universal requirement for voter-verified paper records of all ballots (VVPR). We also need to pass laws that enable candidates to obtain manual recounts easily and inexpensively. There is now a national movement to make sure this technology is used, and it's winning, slowly but surely. Since the 2004 election, state after state has passed laws requirement VVPRs, and others have required VVPRs by administrative decree. In most states, this is the result of grassroots activism by citizens groups with support from national groups. A recent example of an outstanding success is New Mexico's law requiring paper ballots, marked by the voters, which was signed March 2.
There remains much to do on the technology front, including converting hard-core e-voting states like Texas, Florida, and Georgia to VVPR (or, better, adopting a nationwide law, such as the one proposed by Rep. Rush Holt in the House or the one proposed by Sen. John Ensign). Also, the system for certifying voting technology at the state and national level is completely broken. But these problems can be solved with time and dedicated activism.
Election procedures need much greater attention, from the storage of equipment before an election to the storage of ballots after the last recount. Currently, inadequately tested voting machines break down on Election Day. Uncertified and sometimes buggy software is routinely used. Votes are counted behind closed doors. Machines and ballots are in the custody of a single individual, and are sometimes misplaced. Recounts are conducted with rules that are often made up on the spot.
Detailed election procedures need to be defined, taking into account the differences between jurisdictions (including differences in technology). These procedures need to be followed rigorously, even in remote locations with underfunded and understaffed election offices. Procedures need to be improved from election to election, and experiences with new procedures need to be shared among different election offices.
Many of the reforms in technology and procedures need to be codified in election law, including requiring VVPRs. There should be a law requiring the mandatory auditing of election results by manually counting paper ballots from a random sample of the precincts. Routine manual audits depoliticize recounts because they do not have to be requested by a candidate, and because they must occur regardless of whether an election is close or which candidate won. With routine audits, election problems can be discovered and addressed when the outcome of the election is not in dispute.
It is critical that candidates (or, even better, members of the public) be able to obtain manual recounts easily and inexpensively. In recent years, putative winners of close elections have often alluded the "chaos in Florida 2000" for the purpose of suppressing a recount. Recounts conducted under clear rules would not be so chaotic. It is simple common sense to take a close look at the ballots when there is a question about an election. A little cost or effort to satisfy a disgruntled candidate (and his or her supporters) pays huge dividends for democracy.
Finally, these improvements will have little effect unless citizens are more involved in elections. Citizens have to generate grassroots pressure for reforms. There need to be observers to take advantage of any increased openness in election procedures. Indeed, many procedural improvements depend on the presence of independent witnesses to be effective. Citizens need to be see what procedures are actually followed in an election, and compare that with the procedures that should be followed. We have seen time and time again that election laws are routinely ignored—unless someone is watching.
Many of our current problems stem from a "quick-fix" attitude - leading to fresh problems, such as the idea that new touch-screen machines would solve all our election woes. To have the kind of elections we need will take hard work and many years, and there will be setbacks along the way. But if we follow a long-term plan, we'll see that each election is better than the previous.