By Susan Jameson
Granite Bay View
Ah, the dangling chad.
Who'd have thought that a system implemented in the 1960s could cause so many problems in the 2000 national election and subsequent laws and resolutions, including the Help America Vote Act of 2002?
Perhaps Granite Bay programmer Alan Dechert, co-founder and president of the Open Voting Consortium (OVC). The group of citizens and master programmers is dedicated to voting transparency by developing, maintaining and delivering accurate and verifiable open-source software to be used in public voting.
According to Dechert, the problems that drew national attention in the 2000 election were bound to happen, but worse has been the nation's answer to that problem - the direct recording electronic (DRE) paperless voting machine.
"The concept of invisible ballots created with secret software is fundamentally flawed," he says.
Although DREs are voter friendly they are both paperless and use secret programming code to tabulate votes, making it virtually impossible to completely trust the results.
Dechert says that voting machines need to use free, open-source software to be completely reliable. And he is not the only one.
Many voting organizations, computer programmers, engineers and scientists have also spoken out against the paperless system, helping to get legislation passed in California to, at the very least, add a paper trail in the event of a close or contested election. In fact a brand new law (SB370, introduced by Los Angeles area Senator Debora Bowen), requires a voter verified paper trail to be used in state required auditing and contested recounts.
But Dechert and OVC say that is not enough.
"The overall problem is that much of the system is not comprehensible to the average person. The counting of votes is supposed to be public," says Dechert. "You cannot observe tabulation with automated systems as they exist now. We can get a level of transparency, by making the source codes (or the instructions to the computer) public."
Five years ago, Dechert came up with the idea to develop a voting software that will not only print a ballot in the booth for each voter to check, but have code that is in plain language so that any programmer can see how votes are tabulated. The process, according to Dechert, would not only bring voter confidence back up, but, because it can be installed on any computer, could save the government thousands of dollars.
At the time, Dechert was a consultant to the Sacramento County Public Works Department. After an unsuccessful try to get a pilot program going in Sacramento County, he eventually co-founded OVC with University of California, Santa Cruz IT professor Arthur Keller and University of Iowa computer science professor Doug Jones.
By 2003, a team of volunteer scientists and engineers developed a prototype. On April 1, 2004, the group demonstrated the OVC system at the County Government Center in San Jose to local election officials, politicians and the general public. The demo was covered by local media and seemed to be well-received.
Average folks can log on to the group's Web site (www.openvoting.org), click on "Web Demo" and check out the process for themselves.
Basically, voters, using a mouse or a touch screen, select their choices for each office. The program will only allow the maximum number of candidates to be selected and allows for instant run-off selection of local officials through a ranking system. Once completed, the voter prints a sample ballot, which is an exact replica of what they saw on screen on plain paper. The ballot also has a scan code on the side, which is readable by vote-counting machines.
The voter is able to visually check their vote, place it in a folder and then put it in the ballot box. Although the ballot has a tracking number, it is randomly generated so that even in low traffic polls, the ballot cannot be tracked back to the voter.
Reading-impaired voters are able to plug in ear phones to complete their vote and, using a hand scanner and ear phones, are able to check the vote without assistance.
OVC was incorporated in late 2003 and has been a growing force in voting reform. With 25 founding members, seven board members, two full-time employees and lots of volunteers, the consortium now has more than 1,000 contributors. While OVC is known to many government officials nationwide, the OVC system is not certified and not ready for use in production.
"It seems we have to sell the whole country on the idea before we get the necessary funding support to build a production system," says Dechert.
On Oct. 21, the U.S. General Accountability Office (GAO) issued their report (GAO-05-956) to Congress on the security of electronic voting systems. The GAO report lists OVC as a key initiative and says that it "developed a prototype for an open-source electronic voting application that uses commercial hardware and operating system components and provides (1) an electronic voting machine that prints a paper ballot, (2) a ballot verification station that scans the paper ballot and lets the voter hear the selections, and (3) an application to tally the paper ballots."
OVC is cited first in the GAO report along with Caltech/MIT, National Academy of Sciences, the Brennan Center of Justice (New York University) and others in the section on nongovernmental voting initiatives.
OVC also has a working relationship with A Center for Correct, Usable, Reliable, Auditable, and Transparent Elections (ACCURATE), which on Aug. 15 received a five-year, $7.5 million National Science Foundation grant to assess the current voting system. ACCURATE consists of a number of university professors, including computer and behavioral science experts and law professors, from six diverse institutions from all over the country. They are lead by Johns Hopkins University's Avi Rubin. In addition to Johns Hopkins University, UC Berkeley, Rice University, Stanford University, the University of Iowa and SRI International are also part of ACCURATE.
"We're aligned philosophically," says Dechert, who is a Berkeley grad. "And that's what public officials need. They need someone with that kind of reputation to tell them that (open-source voting is the best choice)."
While working for Sacramento County, Dechert developed a software toolkit for correcting applications with Y2K problems, which he was allowed to release to the public domain so that the private sector could use. He got responses from governments and businesses from around the world, including the Department of Defense. He also started a consulting business on the side to help smaller companies use the program.
The open-source model is similar. The idea is that you give away the program and then sell the support services to go along with it.
A New York and Maryland based company is already taking that formula to the East Coast. Open Voting Solutions hopes to offer their open-source voting machine to the East Coast by December. In partnership with major hardware and software producers, like HP and Adobe, Open Voting Solutions will be able to build better machines from off the shelf parts for about 75 percent less money than other electronic vote companies.
"We are focused on building the best, cheapest and most open product," says Open Voting Solutions CEO Richard C. Johnson, who has worked closely with OVC.
Johnson says certification is the biggest, most expensive hurdle for any voting manufacturer.
Dechert is inclined to agree.
"The certification costs are in the six figures," says Dechert. "We need to go to the testing labs and go through all the processes and pay all the fees. They charge $125 an hour and it's probably a six month process."
He is also in support of Johnson and his Open Voting Solutions efforts.
"Open Voting Consortium is not here to sell voting systems. What we are designed to be is an industry supported consortium of organizations that deliver those goods and services. Open Voting Solutions is potentially one of those."
So far, only individual dues paying members have provided financial support to OVC. In the long run, however, it is designed to be supported by larger organizations and companies with an interest in the voting system or open-source software - companies like Red Hat. Recently, Greg DeKoenigsberg, community relations manager of Red Hat, wrote "our long-term mission is to further the democratization of code and content. Open voting systems are an obvious fit for our mission." OVC and Red Hat are currently exploring a relationship, which is expected to be finalized soon.
Right now, Dechert and volunteers like Bay Area real estate broker Brent Turner are busy meeting with state legislators and county representatives to spread the word about their project.
Turner got involved with OVC a little over a year ago when he came across information at a voting rights gathering. After doing his own research on the group, he joined and became an active volunteer.
"I do believe this is a group that has pure intentions," he says. "They want to make sure that vote counts are transparent and can be verified."
Following the San Jose demonstration, Turner took the initiative to call all the county registrars, sent them a DVD of the demonstration and set up appointments for visitations. He says that while Dechert is concentrating on the state capital, other representatives are informing the county registrars, who are often the ones making the decision as to which voting system to use.
Turner and Dechert both agree that they would like the system implemented on a statewide level and eventually move to a national level as states see how well it works in California.
"I hear it all the time, people are looking to California for leadership in this area," says Dechert. "So, we think that if we can get it established in California, then it will spread quickly."
Another benefit to open code software is that because anyone with computer programming knowledge can understand how it works, they can also recommend changes to make it better. This utilizes a country full of computer experts instead of just those on the clock for a certain company. But it doesn't comprise the integrity of the software as it is utilized at the polls.
Dechert likens the process to a car dealer making the design plans for a vehicle available to the public. An engineer shopping for a car could ask to see the plans, see potential flaws and make recommendations on altering future models. His recommendations would not affect the current vehicle on display in the showroom, but if implemented, could affect the design of next year's model.
This system of many minds make a better project was illustrated in 2003, according to an article in Wired, when Diebold (a DRE manufacturer) accidentally revealed its source code and several independent analysts found the programming to be full of serious flaws. With open-source codes, these issues will constantly be undergoing analysis and revisions to make the system more accurate and secure.
"This is a bi-partisan issue," says Turner. "Everyone should want the votes counted accurately."
And on the surface everyone does. Who would oppose a system that counts the votes openly and accurately and is cheaper to implement?
"The implications of a successful project for us are huge for the whole open-source community," says Dechert. "When I testified a year ago in support of (open-source voting), the opposition was not from the voting machine vendors. It was all from industry."
Companies, like Microsoft, who profit from proprietary software, see free open-source software as a threat. And they should. According to annual revenue reports from Microsoft, IBM Software and Red Hat (leading commercial vendors of Linux and other open-source software), Red Hat's annual growth percentage has been growing rapidly over the last two years, while IBM and Microsoft have remained steady.
According to Dechert, corporations like Microsoft see open-source voting software as a huge leap in the acceptance of open-source software as a whole. If the government embraces open-source, the public is soon to follow. Proprietary software, like Windows, Word and Office that need consumers to purchase upgrades often would be replaced by free, open-source operating systems like Linux, leaving corporations who base profits on these upgrades in the dust.
John Miché of Orangevale, who is a supporting member of OVC, says he is involved because his company, Bento Solutions, has connections with open-source software.
"I really see open-source as the coming element in software," says Miché.
Last year, with the help of Assemblywoman Jackie Goldberg, OVC got Assembly Concurrent Resolution 242 (ACR 242) introduced and passed in the State Legislature. ACR 242 calls on the Secretary of State to investigate using open-source software for elections and issue a report on it by the start of next year.
When California Secretary of State Bruce McPherson took office in the spring it became his responsibility to create that report. Originally a panel of experts was to be created to study the subject and create the report. But the implementation of the panel has been slow in organizing, despite a Jan. 1, 2006 deadline for its findings.
"The whole thing has not been completely decided yet," says Dechert. "But if we are going to get it done in a timely fashion, the whole thing does really need to get started."
Sacramento Senator Deborah Ortiz has urged McPherson to not only form the panel but also to hold public hearings on the issue.
In a letter dated Oct. 11 to McPherson, Ortiz states, "To guarantee confidence in our government institutions and decisions, it is important that the public have access to the work of task forces and working groups that inform agency decisions."
With any luck the panel will be implemented before the end of the year and Dechert and his group can move on to the how of the matter instead of spreading the word as to why it is needed. Dechert would be pleased to see open-source software at the heart of California's elections in 2006.
And he has many voters behind him.
Susan Jameson can be reached at firstname.lastname@example.org.