Like many startups that aim to do something big, Open Source Digital Voting Foundation in Palo Alto began with an "Aha!"
The epiphany belonged to Gregory Miller, 50, a technology marketing executive who earned a law degree in intellectual property and is now Digital Voting's chief development officer.
"Back in November 2006 I was entrepreneur-in-residence at a local venture-capital firm. The week of the election I happened to have a conversation with one of the partners about what we called 'malformed' markets, and we naturally turned to a discussion of the voting-systems industry, which was failing on all fronts at the time -- bad products, dysfunctional business models, the works," Miller said.
"It's a lousy business if you want to make money for your investors and do a good job for voters."
Miller turned to friend and colleague John Sebes, 47, a Silicon Valley veteran with a background in computing infrastructure and security.
"He dropped by the house one afternoon and started telling me about how messed up voting machines are," said Sebes, who is now the organization's chief technology officer. "I was naive enough to think it would be a simple technical fix and signed on."
Located in downtown Palo Alto, Open Source Digital Voting Foundation's goal is to build a publicly owned digital elections system that is practical, secure, affordable and above reproach.
It would include voter registration, ballot design, voting, and counting and reporting of election results.
"Re-inventing How America Votes in a Digital Democracy" is the nonprofit organization's slogan.
"Our mission is to make digital voting and election automation believably honest, even under intense scrutiny in the closest, most contested races," said Sebes, while sitting in his Lytton Avenue office recently.
"There are three requirements for this. First, ensure to the greatest extent possible that only those entitled to vote can successfully do so, minimizing potential fraud," he said.
"Second, make registration and voting as accessible, inclusive and easy as possible, maximizing participation in our democratic process.
"And lastly, make the casting and counting of ballots completely open to public examination, auditable and therefore verifiably accurate."
He paused, smiled and then deadpanned, "It's that simple, really."
Sebes and Miller, who have known each other for more than 15 years, started by immersing themselves in the fundamentals for legal, accurate and efficient elections, not just in California but across the nation.
"Along the way we established key partnerships with state secretaries and elections directors, various taskforces, public policy and interest groups and, of course, other technologists," said Sebes, a Zen gardener and aspiring trumpet player in his spare time.
This exercise convinced them the software programs that drive digital elections should be "open source" and emphasize enterprise-computing practices that have been proven to have high integrity in the private sector.
And to be successful, the system must be designed from the start to enable incremental adoption and easy adaptation, and it must win the public trust, contrasting with the widespread skepticism about existing commercial voting systems.
"Once we got a handle on the issues, we knew our focus should be on building things," Sebes said.
"Our job is not to be a think-tank or an advocate for election reform. We are a hands-on, nonprofit developer of election software."
This realistic, "fix what is there" utilitarian view attracted initial funding from Mitch Kapor, the San Francisco-based entrepreneur, philanthropist and founder of Lotus Development, a software powerhouse in the salad days of personal computing. Kapor also introduced Sebes and Miller to Aleks Totic and Pito Salas, both Internet and personal-computing veterans, to round out the bench of managers needed to move the startup from ideas to action.
The Foundation is now in its third year of a plan that may take up to eight years to complete. It has a technology "roadmap" in place and is actively engaged with elections officials in eight states.
Its flagship project, "Trust the Vote," is well underway, with its first software product -- a voter-registration platform -- already in use by the largest non-partisan voter-advocacy group in the country.
Across the country, when most voter registration forms are processed or election ballots counted, a computer system is somehow involved, according to California Secretary of State Debra Bowen.
This is true even in small states such as New Hampshire, which has fewer than 760,000 registered voters compared to California's estimated 14.9 million.
Bowen traveled to New Hampshire on Election Day in 2008 to investigate whether all ballots were still counted by hand.
"I watched," Bowen said. "In most of the places I'd been told that they hand-counted all the ballots they actually didn't. They only hand-counted one, maybe two races, and the rest they counted by computer."
The use of computers in election processes caused little discernable public concern until the 2000 national election, when George W. Bush edged out Al Gore for the presidency in a hairline vote in Florida. Nearly a decade later the bitter legacy of hanging chads, butterfly ballots and a split Supreme Court deciding who would occupy the White House still lingers.
Despite the Help America Vote Act passed by Congress in 2002 and billions of dollars spent to improve voting technology, trust-eroding problems still occur. The 2004 national election, narrowly won in a handful of precincts dotting battleground states like Ohio, was dogged by controversy.
Just two years later, voting-machine failures were widely reported during the 2006 midterm election. Bowen ordered a top-to-bottom review and recertification of all voting machines in California. She decertified many of them while ordering stricter procedures for those remaining.
Although the 2008 Presidential election was less controversial, Sebes and Miller remain adamant that problems with digital voting have not gone away.
"Greater margins of victory in 2008 undoubtedly helped," Miller said. "But that, together with the fact that a lot of expensive voting machines got thrown out, only serves to obscure serious vulnerabilities that continue to plague the entire elections process.
"In our view the trend looks pretty clear -- there likely will be more, not fewer, highly contentious elections going forward. And where victories are paper thin, there are likely to be problems."
Miller asserts that the nation's elections infrastructure hasn't improved significantly since the 2000 election.
"It's scary when you think about it. It's nothing less than a political and civic mid-air collision of the first order, if not an outright constitutional crisis, just waiting to happen."
Bowen supports the continued use of paper ballots, particularly as source documents when detailed audits are needed. But she says computers are also a necessary tool for conducting elections.
"I think at this point it's unrealistic to expect in California that Dean Logan (Los Angeles County registrar-recorder) and his staff are going to find a way to hand count 4 million ballots with 120 races each," Bowen said.
"I think we need to take advantage of the power of the computer where we can, but back it up with systems that tell us when we might have made a mistake so we don't make it an irreparable mistake."
While the Digital Voting Foundation philosophy is that there is a role for automation in election operations, Sebes said a new digital system is no cure-all.
The paramount objective of elections reform goes deeper than automation, he said.
"There are plenty of things about elections that need thoughtful attention and improvement. But the really important ones -- and probably the hardest ones -- don't have anything to do with computers; notable among them are laws and procedures in need of reform. In the meantime, we want to take things as they are and make them better.
"To be blunt about it, we mostly want to take computers and technology out of the mix of finger-pointing and cynical blame-gaming that only serve to turn people away and undermine our most important rights and responsibilities as free people."
A key premise of the Digital Voting Foundation is that elections software should be developed and maintained using an "open source" approach. But what is open-source software?
Webopedia, an online encyclopedia of computer technology, defines it as "a program in which the source code is available to the general public for use and/or modification from its original design free of charge, i.e, open. Open source code is typically created as a collaborative effort in which programmers improve upon the code and share the changes within the community."
Since the coding is available for review at all times and contains no trade secrets, there is nothing that is hidden from view and therefore unknown or unknowable.
"There is a saying in our (high technology) industry about open source," Miller said. "That when a thousand eyes are looking all bugs are shallow. And what that means is that ... you can't hide anything, there are no trap doors, there are no secret ways in, there are no easily modified things while no one is looking."
Advocates of open source contend the process leads to better software, that needed fixes and improvements are accelerated and that it is far more difficult, if not impossible, to add malicious capabilities that go undiscovered for long.
Bowen, who is widely respected in the Valley for her technology savvy, acknowledges the benefits -- but adds that open-source software cannot, in and of itself, guarantee fair and unfettered elections.
Voters shouldn't trust it any more than they would vendor proprietary products "just because some computer experts say that type of system is fine," she said.
But "it certainly could go a long way in election transparency."
Sebes agrees: Transparency is the attribute most valued by the Digital Voting Foundation staff in its quest to make elections trustworthy.
"We are not open-source zealots. In this case it just makes good, practical sense."
Fixing and improving the patchwork of information technology used to register voters, administer elections and count votes is anything but straightforward.
For example, Los Angeles County, at 4,082 square miles, is the largest voting jurisdiction in the United States. With more than 4.3 million registered voters in the county -- a number greater than that of 25 individual states and Puerto Rico -- it is also one of the most diverse and complex. There are 88 cities in the county with 4,394 polling places. Twenty-five different elections were held between 2006 through 2009 for which ballots were provided in seven languages (English, Chinese, Japanese, Korean, Spanish, Tagalog and Vietnamese) as well as assistance for three others (Armenian, Cambodian, and Russian).
In contrast, California's smallest county -- Alpine -- lists 799 registered voters on its rolls. Santa Clara County has 789,175 voters. Thus, there is no workable "one size fits all" approach to elections automation.
The Digital Voting team chose voter registration as the place to start for its first production software release. Among all electoral activities, voter registration is one of the most debated, partisan and controversial.
Loose procedures can lead to voter fraud, allowing illegal immigrants to cast ballots, for example. Yet onerous registration procedures can result in voter suppression by making it overly difficult for the poor, the young and others to vote regularly or at all.
Overlying this debate is the reality that voter registration -- especially in densely populated areas -- imposes significant burdens on understaffed elections offices that must process paper applications.
The paper burden reaches to the top: "If you were to walk into my office about three weeks before the voter registration deadline (you would) see all of us, including me, sorting out voter registration forms to send to the 58 counties by hand," Bowen said.
As many as 60,000 registration cards arrive daily at Bowen's office during that time.
Processing the paper registration forms is a challenge.
In California, it usually goes something like this: Each form is manually inspected at the registrar's office of the county where the applicant resides. When a form is rejected, its applicant is notified by mail and invited to try again.
Information contained in registrations that are accepted is manually typed into one or more computerized systems that manage voter rolls. Each county currently maintains its own list but eventually the state will provide a centralized system.
Bowen has said that data entry, like any repetitive task, is prone to error, especially during peak periods.
The result arguably encompasses the worst of two worlds, according to the Digital Voting staff: For would-be voters, the manual-registration process can be slow and unforgiving; for elections officers, the process is labor-intensive and error-prone.
"This is precisely the kind of conundrum we are trying to address," Sebes said. "We're happy to embrace the need for paper documents, but want to make them easier to fill out and, where possible, eliminate manual data entry," while allowing for local adaptations and incremental, one-step-at-a-time implementation, which government agencies want and need.
In Digital Voting's registration system, a "wizard" guides would-be voters to fill out a form online, which can then be printed out. The program prevents applicants from printing registration forms that are incomplete or contain errors. This alone improves the quality of applications arriving at elections offices, thus reducing workloads and improving accuracy, Sebes and Miller assert.
The Digital Voting design can also attach a unique barcode to and capture the data in each application at the time it is printed, which makes it additionally trackable.
In one scenario, an elections clerk could receive an original, signed application. Instead of manually entering the information into the local voter-management system, the clerk would merely call up the digital copy by scanning in its barcode. The signed original would then be checked against its image. If there is a match, approval can be completed with a single keystroke.
This speeds the approval process and reduces errors and costs, according to Digital Voting.
Also, with barcodes, audits could analyze data that would reveal whether an individual clerk or other "insider" appears to reject a disproportionate number of voters who live in a given neighborhood or are affiliated with a particular political party, for example.
For California, Digital Voting's registration software would not only be freely available and open, it also would accommodate signed paper applications as required by current law, making them easier to manage and audit. Sebes believes the system would offer "the best of both worlds, automation and paper," conserving scarce staff and fiscal resources while building confidence in the registration process and elections as a whole.
One of the largest nonprofit organizations involved in voter registration deployed Digital Voting's software in 2009. RockTheVote, which focuses on younger voters, is the single largest third-party registrar in the country, according to Executive Director Heather Smith.
During the 2008 election cycle 2.6 million citizens registered directly or indirectly through Rock the Vote's online services, which it currently hosts across about 23,000 partner sites -- including blogs, Facebook pages and MySpace profiles.
Since going live last September, 40,000 people have used the Rock the Vote online system to register to vote. Sebes said he expects to see a larger spike this summer.
Digital Voting is actively working with eight states -- California, New Hampshire, New York, North Dakota, Ohio, Oregon, Vermont and Washington -- collaborating with some on specific Trust the Vote components. More than 25 percent of the country's voters reside in these states.
It is also in conversations with officials in another 11 states, including Michigan and Texas -- a group encompassing another 25 percent of the nation's voters.
The foundation's leaders hope that by demonstrating the capabilities of its voter-registration program there might be acceptance of open-source software at the federal level, as well as certification of Digital Voting's software in upcoming national elections. Much of the work is related to two key pieces of Congressional legislation: The 1986 Uniformed and Overseas Citizens Absentee Voting Act and its follow-on bill, the 2009 Military and Overseas Voter Empowerment Act.
These laws are intended to streamline registration and voting for the 6 million Americans -- military and civilian -- deployed or residing abroad. They require the states to make online registration available to them.
"Near-term we hope to work with some of those states to help build the capabilities they need to conform with overseas and expatriate voter requirements," Sebes said. "We're very excited about what that means in terms of motivating broad support for putting reasonable and open digital-elections solutions in place going forward."
Sebes and Miller say they already have partnerships in place, an understanding of elections requirements, software that has been deployed and blueprints for the future. Much more work lies ahead.
Digital Voting needs to broaden its financial support and attract an ever-wider community of volunteer contributors. And without a critical mass of technologist and other contributors, it risks failure, like so many startups, despite good intentions.
But Sebes and Miller are hopeful their venture will bring about the large-scale change they've envisioned.
"Once the voting systems in place are broadly viewed as trustworthy it's more likely that a serious national conversation -- not just another polarizing food fight -- will emerge next time we have a big problem, which in our view is highly likely to occur," Sebes said.
"The conversation we need is about larger, more salient issues, like arcane statutes and electoral processes, all in urgent need of calm discussion and sensible reform."