this is Alan Dechert's talk given at the UC Santa Cruz forum on electronic voting held October 26th, 2003

Good afternoon. I thank all of you for coming. I am glad to be here. Thanks also to Bob Kibrick and Arthur Keller for their work in organizing this forum. I also want to say thank you to all the others at UC Santa Cruz that helped make this possible, and to all the other forum speakers.

Ladies and Gentlemen, we have a problem. Our voting system is broken. It does not need to be repaired, however. It needs to be replaced. We don’t need to upgrade our voting system. We need to replace it. We don’t just need some new voting machines. We need a whole new idea for how to administer public elections.

This afternoon, I will describe the work some of us are doing to bring about a better system. But first, I want to take a few minutes to set the context.

Sovereignty is the right to decide. Sovereignty means the right to decide.

In a democracy, sovereignty is … distributed. In a democracy, sovereignty is distributed, while you — the individual citizen — retain most of this right to decide.

We share this right. We distribute sovereignty. The system for distributing sovereignty is called democracy. In a democracy, we distribute sovereignty through our system of laws. The voting system is part of the system whereby we share our right to decide. Sometimes, we vote to decide on specific issues. But mostly we vote for people who will make decisions for us.

While democracies all share the rule-of-law concept, they vary widely on voting systems. In the U.S., it seems that we vote on everything. We vote for everything from President to dogcatcher. We vote on a lot of initiatives.

We tend to consolidate the administration of elections. While we’re electing federal officials, we say, "Let’s put state and local contests on the same ballot with them." This is very unusual. Only the Swiss do something similar. In most democracies, you vote for your Member of Parliament and that’s it, at the national level. Local elections are separate and may be administered — as in Canada — with different rules and different equipment by different entities.

Is our system more democratic because we put more things on the ballot? Not necessarily. It is not practical to present every issue to the people for vote — no government even attempts to do that. The vast majority of political jobs are still appointed. Something like 2,000 jobs were at stake in the recent contest for Governor of California. Democracy is not proportional to the number of offices and issues on which the people vote.

Do we really need so many elected offices, anyway? An idea going back to ancient Greece suggests maybe not. Most officials were chosen by sortition: They were chosen by lottery from a pool of qualified individuals.

Our voting system varies within the United States. Voting systems vary from state-to-state and county-to-county. Warren Slocum is the chief of elections for San Mateo County. He was elected: most of his counter- parts in other counties were appointed.

In January of 2001, I attended a hearing conducted by the State Assembly’s Elections committee. Some of the testimony was incredible. One of the many horror stories had to do with a man that was not registered to vote where he lived on one Election Day, and was given a provisional ballot. For some reason, he seemed to like voting with the provisional ballot. So, rather than re-registering, year-after-year he voted a provisional ballot. Only one problem: his votes were never counted. We can only speculate why he didn’t want to be in the normal voter file; but sometimes voter lists have been used for marketing, and they are also used to find people for jury duty. There are several reasons someone might like not to be on the list but still be able to vote. Anyway, after a few years, the elections people recognized his signature by sight but it was against their policy to inform owners of provisional ballots whether or not their vote was accepted. Year after year, his vote was thrown out. Now, how dumb is this system?

At least in Iowa, if you go to the polls to vote and you’re not on the roster, they will give you a provisional ballot but they will also make sure you get registered to vote at the same time.

Voters should not be disenfranchised by the voting system. By voter disenfranchisement I am talking about cases where votes are not counted the way the voter intended, or the vote was not counted at all, or the voter was discouraged from voting altogether — even though registered to vote. Furthermore, I am talking about cases where potential voters were discouraged from registering to vote.

If we’re going to have voter registration — and it’s not clear to me that we need it — it should not be an obstacle for voters. But currently, it is an obstacle.

In Georgia, in order to vote you have register by the 4th Friday before Election Day. In Ohio, the deadline is 30 days. In Iowa, it’s 10 days before any Primary or General election, but 11 days for other elections. It’s the 3rd Saturday before in Vermont, but you only have until noon to get your paperwork into the town clerk. In North Dakota, they do not have voter registration. In Idaho, the deadline is 25 days before Election Day but then you can register on Election Day itself. What if you moved from Idaho to California a week or two before Election Day? You might think you would still be able to register on Election Day. You would be out of luck. Rarely an issue, you say? Consider that almost one million Americans move in any given week. Over a million voters could be impacted by the logistics of moving while having "register-to-vote" on a long to-do list.

At a minimum, Election Day registration should be universal. In the long run, we should investigate how we can eliminate voter registration altogether. Afterall, if you are a citizen of voting age and you meet all the requirements, you are also in many databases. Why have another database that has serious maintenance issues for the voters as well as state and county governments?

As I see it, voter registration, provisional ballots, and a whole lot of the other rigmarole and gobbledygook are artifacts of the 20th century — before the Internet and before the personal computer. In this new century, every polling place should be securely connected to the Internet. Studies have shown that while you will probably never be able to vote remotely over the Internet unattended, it should be possible to vote over the Internet at poll sites where your identity can be confirmed. We should look at this as a possible replacement for current absentee ballot methods.

In the wake of the election mess in 2000, the presidents of Caltech and MIT launched a project they hoped would bring about a solution to the voting system… conundrum. Their December 14th 2000 press release was titled, "Caltech and MIT Join Forces to Create Reliable, Uniform Voting System." Note the word, "uniform." In other words, they noticed — like most of us — the great variety of faulty procedures. Then they said, "Let’s create new voting technology, test it thoroughly…make it bullet proof … then use the same equipment and procedures everywhere." Why not have a uniform system? Why have all these daffy different ways of doing the same thing? Why not find one way that really works and use that?

Quoting the beginning of the press release,

The presidents of MIT and Caltech have announced a collaborative project to develop an easy-to-use, reliable, affordable and secure United States voting machine that will prevent a recurrence of the problems that threatened the 2000 presidential election….

Okay, so they wanted to develop a U.S. voting machine. It sounds like they started with a really good idea. But now, almost three years later, where’s their voting machine? Where is this technology they were talking about? This is not rocket science, folks. Maybe that’s the problem. Caltech and MIT are good at rocket science. They know how to do that. But the obstacles here are mainly political. The voting system represents a significant technical challenge, but it’s mainly applied science. Technically, this is not horribly difficult. Politically, it is a pain.

Of course, I think they had the right idea. In fact, I published a similar idea about the same time as their announcement. Before you ask, "Okay, now, three years later, where’s Alan Dechert’s voting machine?", let me make this point: I am critical of Caltech and MIT researchers because they started with a lot of money and tremendous institutional support. They did not follow through with the idea. I started with none of these resources. Maybe you will say, "He’s just envious." And maybe that’s it: I’m just envious of their resources. But the fact is that when it comes to the idea of a uniform system and a U.S. voting machine, I have moved the idea forward while they have not. I also incorporated the open source concept and the voter-verified printed ballot. Caltech/MIT never embraced these ideas. They abandoned the idea of a uniform system and chose to work with existing vendors in effort to improve voting systems.

There is still a Caltech/MIT voting project and they are doing some good work — uncovering information and helping to improve things here and there. But it is not the magnificent work they set out to accomplish.

I am developing the Open Voting Consortium as a durable organization to develop, transfer, and maintain this new voting technology. There is no need for expensive dedicated hardware: we can use inexpensive off-the-shelf commodity PCs and peripherals, while utilizing free software. Once our open voting system becomes established, the persistent issue of how to replace obsolete hardware will be gone forever. Usually, we’re happy to pay a little more to get something better. It happens that our system will be much better, but also much much cheaper.

A secure, reliable, trustworthy, and affordable voting system is an essential feature of a successful democracy. A voter-verified paper ballot is an essential feature of a successful voting system.

To give you an idea of what your ballot may look like in the future, I’ve passed out some samples in folders like this (hold up sample). With our system, you print the ballot yourself on ordinary eight-and-a-half by eleven paper in the voting booth, and then place it in a privacy folder like this. You’ll notice about half an inch of the ballot is exposed. The exposed border has the barcode, but no other printing.

This part of our demonstration system — the ballot printing function — is done. It’s on the Internet, and available for anyone try out.

This is what we mean by a voter-verified printed ballot. You go to the computer; make your selections; print it out with the touch of a button, and look to verify that these are the selections you intended to make. If not, you can destroy this one and start over. No problem. There are no hanging chads, no ambiguous marks. No voter intent issues — it’s all written out very clearly.

What if because of some disability you can’t read the ballot? First of all, you would have voted at a station where all the selections were presented orally through headphones, much like the way existing electronic systems work. Unlike existing systems, however, you would have an opportunity to actually verify how your vote was recorded. It is not possible to verify your vote on existing electronic voting machines — visually or otherwise. Anyone claiming voter verifiability for these paperless systems is telling fairy tales. At best, all you can get is some indication that the machine knows what selections you want to make. This may or may not have anything to do with how your vote gets recorded. With our system, even if you can’t read it, you will print the ballot just like everyone else. If you want to verify your ballot, you can go to a station with a scanner, and have the barcode scanned while you’re wearing headphones so you can hear your selections read to you. The ballot does not have to be removed from the folder: Your privacy is assured. Then you go to the ballot box and deposit your ballot like everyone else.

Right now, we also have the software available for download that will take the encoded information from the barcode and read back the selections. If you want to try that out, it’s available. We’ll have a full demonstration of our system ready in a few weeks.

In summary, while technology has taken great strides forward in recent decades, the voting system has not kept up. The problem of modernizing the system has been routinely underestimated. It’s very complicated.

The technical complexity is routinely underestimated, but we know how to do it. It is the political complexity that boggles the mind and tests our resolve. But our team is meeting this challenge!

THANK YOU.