Recapping The OSCON O’Reilly Radar Conversation
A couple of weeks ago I presented at OSCON and during the conference had an opportunity to sit down with Mac Slocum, Managing Editor for the O’Reilly Radar. We had about a half an hour conversation, for which we covered ~20 minutes of it on camera. You can find it here if you want to watch me jaw. But perhaps simpler below, I’ve listened to the tape, and captured the essence of my answers to Mac’s questions about what the Foundation is about and working on and the like. I promised Matt Douglass, our Public Relations Director I’d get this up for interested followers; apologize it took me a couple of weeks.
So, here it is; again not an official transcript, but a compilation of my answers after watching and listening to the video interview about a dozen times (so you don’t have to) combined with my recollection as close as I recall my remarks – expressed and intended.
O’Reilly: How are voting systems in the U.S. currently handled? In other words, where do they come from; procurement process; who decides/buys; etc.?
Miller: Voting systems are currently developed and delivered by proprietary systems vendors, and procured by local election jurisdictions such counties and townships. The States’ role is to approve specific products for procurement, often requiring products to have completed a Federal certification process overseen by the EAC. However, the counties and local elections jurisdictions make the vast majority of elections equipment acquisition decisions across the country.
O’Reilly: So how many vendors are there? Or maybe more to the point, what’s the state of the industry; who are the players; and what’s the innovation opportunity, etc.?
Miller: Most of the U.S. market is currently served by just 3 vendors. You know, as we sit here today, just two vendors control some 88% of America’s voting systems infrastructure, and one of them has a white-knuckled grip on 75% of that. Election Systems and Services is the largest, after having acquired Premier Systems from its parent company, Diebold. The DoJ interceded on that acquisition under a mandatory Hart-Scott-Rodino Act review to consider potential anti-trust issues. In their settlement with ES&S, the Company dealt off a portion of their technology (and presumably customers) to the Canadian firm Dominion Systems. Dominion was a small player in the U.S. until recently when it acquired those technology assets of Premier (as part of the DoJ acquisition, and acquired the other fomer market force, Sequoia. And that resulted in consolidating approximately 12% of the U.S. market. Most of the remaining U.S. market is served by Hart-Intercivic Systems.
On the one hand, I’d argued that the voting systems marketplace is so dysfunctional and malformed that there is no incentive to innovate, and at worst, there is a perverse disincentive to innovate and therefore really not much opportunity. At least that’s what we really believed when we started the Foundation in November 2006. Seriously, for the most part any discussion about innovation in this market today amounts to a discussion of ensuring spare parts for what’s out there. But really what catalyzed us was the belief that we could inject a new level of opportunity… a new infusion of innovation. So, we believe part of the innovation opportunity is demonstrated by the demise of Premier and Sequoia and now the U.S. elections market is not large or uniform enough to support a healthy eco-system of competition and innovation. So the innovation opportunity is to abandon the proprietary product model, develop new election technology in a public benefits project, and work directly with election officials to determine their actual needs.
O’Reilly: So what is the TrustTheVote Project, and how does that relates to the Foundation?
Miller: The Open Source Digital Voting Foundation is the enabling 501.c.3 public benefits corporation that funds and manages projects to develop innovative, publicly owned open source elections and voting technology. The TrustTheVote Project is the flagship effort of the Foundation to design and develop an entirely new ballot eco-system.
What we’re making is an elections technology framework built on breakthrough innovations in elections administration and management and ballot casting and counting that can restore trust in how America votes. Our design goal is to truly deliver on the four legs of integrity in elections: accuracy, transparency, trust, and security.
The reason we’re doing this is simple: this is the stuff of critical democracy infrastructure – something far too much of a public asset to privatize. We need to deliver what the market has so far failed to deliver. And we want to re-invent that industry – based on a new category of entrants – systems integrators who can take the open source framework, integrate it with qualified commodity hardware, and stand it up for counties and elections jurisdictions across the country.
We’re doing this with a small full time team of very senior technologists and technology business executives, as well as contractors, academia, and volunteer developers.
We’re 4 years into an 8 year undertaking – we believe the full framework will be complete and should be achieving widespread adoption, adaptation, and deployment by the close of 2016 – done right it can impact the national election cycle that year. That said, we’re under some real pressure to expedite this because turns out that a large number of jurisdiction will be looking to replace their current proprietary systems over the next 4 years as well.
O’Reilly: How can open source really improve the voting system?
Miller: Well, open source is not a panacea, but we think it’s an important enabler to any solution for the problems of innovation, transparency, and cost that burden today’s elections. Innovation is enabled by the departure from the proprietary product model, including the use of open-source licensing of software developed in a public benefits project. Transparency, or open-government features and capabilities of voting systems are largely absent and require innovation that the current market does not support. Cost reduction can be enabled by an open-source-based delivery model in which procurements allow system integrators to compete for delivery license-free voting systems, coupled with technical support that lacks the vendor lock-in of current procurements. Open source software doesn’t guarantee any of these benefits, but it does enable them.
I should point out too, that one of our deepest commitments is to elections verification and auditability (sic). And our framework, based on an open standards common data format utilizing a markup language extension to XML called EML is the foundation on which we can deliver that. Likewise, I should point out our framework is predicated on a durable paper ballot of record… although we haven’t talked about the pieces of the framework yet.
O’Reilly: Well our time is limited, but you must know I can’t resist this last question, which is probably controversial but our audience is really curious about. Will online voting ever be viable?
Miller: Well, to be intellectually honest, there are two parts to that loaded question. Let me leave my personal opinion and the position of the Foundation out of it at first, so I just address the question in a sterile light.
First, online voting is already viable in other countries that have these 3 policy features:  a national ID system,  uniform standards for nationwide elections, and  have previously encouraged remote voting by mail rather than in-person voting. These countries also fund the sophisticated centralized IT infrastructure required for online voting, and have accepted the risks of malware and other Internet threats as acceptable parts of nationwide online voting. For a similar approach to be viable in the U.S., those same 3 policy features would likely require some huge political innovations, at the 50-plus state level, if not the Federal level. There really isn’t the political stomach for any of that and particularly national ID although arguably we already have it, or creating national elections and voting standards, let alone building a national elections system infrastructure. In fact, the National Association of State Secretaries recently passed – actually re-upped an earlier resolution to work to sunset the Federal Elections Assistance Commission. In other words, there is a real Federalist sense about elections. So, on this first point of socio-political requirements alone I don’t see it viable any time soon.
But letting our opinion slip into this, the Foundation believes there is a more important barrier from a technical standpoint. There are flat out technical barriers that have to be cleared involving critical security and privacy issues on the edge and at the core of a packet-switched based solution. Furthermore, to build the kind of hardened data center required to transact voting data is far beyond the financial reach of the vast majority of jurisdictions in the country. Another really important point is that online elections are difficult if not impossible to audit or verify. And finally, there is a current lack of sophisticated IT resources in most of the thousands of local elections offices that run elections in the U.S.
So, while elections remain a fundamentally local operation for the foreseeable future, and while funding for elections remains at current levels, and until the technical problems of security and privacy are resolved, nationwide online voting seems unlikely in the U.S.
That said, we should be mindful that the Internet cloud has darkened the doorstep of nearly every aspect of society as we’ve moved from the 2nd age of industrialism to the 3rd age of digitalism. And it seems a bit foolish to assume that the Internet will not impact the conduct of elections in years to come. We know there is a generation out there now who is maturing having never known any way to communicate, find information, shop, or anything other than online. Their phones exist in an always-on society and they expect to be able to do everything they need to interact with their government online. Whether that’s a reasonable expectation I don’t think is the issue.
But I think it will be important for someone to figure out what’s possible in the future – we can’t run and hide from it, but I believe we’re no where near being able to securely and verifiably use the Net for elections. There is some very limited use in military and overseas settings, but it needs to be restricted to venues like that until the integrity issues can be ironed out.
So, we’re not supporters of widespread use of the Internet for voting and we don’t believe it will be viable in the near future on a widespread basis. And honestly, we have too much to do in just improving upon ballot casting and counting devices in a polling place setting to spend too many cycles thinking about how to do this across the Internet.