In addition to the restricted environment for open source business models discussed in the next section, there are also significant regulatory, economic, organizational and perceptional barriers to the use and development of open source software in the voting systems market. In terms of voting system regulation, any changes in system source code trigger system recertification at all levels. Unlike traditional open source software where the ability to change the software frequently is important, open source voting system software development would have to operate differently and take into account that once a product is out on the market, it will be very difficult to change or ``patch'' the software. In addition, federal and most state certification processes are evaluations of an end-to-end system; it will be insufficient to simply develop the software, as any successful certification will have to include hardware, documentation, and procedures in addition to the software.
Even with sufficient attention to planning and development, it will still be difficult for small firms or non-profits to get a foothold in the elections systems market. It takes quite a bit of infrastructure and financial backing to be able to develop, certify, market, implement and service voting systems. Federal certification alone can take from two months to a year and cost between $150,000 and $400,000 for a single voting system.Coggins, C. Independent testing of voting systems. Communications of the ACM 47:10, 34-38 (2004). Contractual performance bonds -- where a vendor puts a certain percentage of the cost of a contract in escrow until the system has performed according to a set of criteria in the contract -- can be hundreds of thousands to millions of dollars. Due to the nature of state and federal voting systems standards and guidelines, voting systems must be certified as end-to-end voting systems -- including precinct-tabulation, data storage and central tabulation -- or a vendor of a subsystem has to team up with a larger firm that has the missing pieces and is willing to sponsor a full system certification.Vogue Election Products & Services, LLC. did just this recently when it teamed up with Election Services and Software to certify and market the AutoMARK ballot marking device.
Of course, other pieces of a voting system business outside of code development need to be in place to field a product. To support the requirements of certifying and marketing an end-to-end system, an open source voting systems vendor will need to have a support organization the likes of which no other open source software applications have had to develop. Some open source businesses such as MySQL AB and SugarCRM do have extensive marketing and support infrastructures for their paying customers, but no open source business produces a product like an end-to-end voting system where software, hardware, documentation and procedures are developed, evaluated, marketed, sold and maintained throughout the lifetime of the product. (Ping says: Huh? Why aren't MySQL, etc. examples of such?)
Finally, in addition to these regulatory, economic and organizational barriers, there are a number of perceptional barriers related to voting system customers that an open source voting system vendor would have to overcome. First, voting system customers -- typically local election officials -- might not understand the debate around disclosure and system security. The intuitive view is that disclosing system source code will result in a less-secure system. Vendors will have to take care to explain the arguments against ``security through obscurity'' and how openly published algorithms, for example in cryptography, have proven more robust to attack. Also, to make a sale, open source vendors will need to be able to demonstrate that the organizational structure they choose will be able to support the system over its lifetime or provide alternatives to such support if the vendor goes out of business.