« previous 1 2 3 4 next »

Slipping card feeds and hanging chads

Superficially, the attractions of using modern technology as part of the electoral process are obvious. In the UK all votes are still entered and counted by hand. In theory a machine-recorded vote would be more reliably assessed, less susceptible to human error, and counts will take a fraction of the time required for counting ballots by traditional means.

In fact, the embarrassment of the US experience in Florida during the November 2000 presidential election, where votes were recorded on punch cards and counted by machine, machines that suffered card slip and 'chad' miscounts, has shown this to be wrong, even where machines are used merely to record a vote. It can be argued, and has been many times, that the failings of voting technology subverted the popular vote in Florida, and changed the outcome of the presidential election.

Many advocates for the application of computer technology to the vote have come to the conclusion that free and open source software offers a cure for these kinds of problems, because the simple expedient of making the code visible enables scrutiny and transparency, and facilitates the exposure and removal of bugs.

There are free and open source voting systems that present a viable alternative to the proprietary model. VoteBox is under development at Rice University, and the Open Voting Consortium demonstrated their model at LinuxWorld last year. The base requirements for the system developed by the Open Voting Consortium are that:

• The whole voting system must be open to complete public scrutiny with no room for "trade secrets"
• The voting system must include a durable paper ballot that can be handled, stacked, counted, and recounted if necessary
• The voting system should invite public participation instead of discouraging it
• Ballots should be cast in private but counted in public
• Accessibility features should be built into the system
• Elections should be conducted in a cost effective manner

For obvious reasons, the authors of the Election Technology Council's report, who have an interest in preserving the status quo, take an opposite view, and believe that open source code offers no advantage to the voting process. The crux of the argument is that "successful open source projects have benefited from operating in an unregulated environment which can fully incorporate a large number of contributors to update and add to the software project" whereas the voting industry is now subject to strict certification and accountability guidelines.

Free and open source software advocates would argue that this assertion is based on a one-dimensional (and mistaken) view of open source development that assumes that all open source projects are developed on the model of individual hackers cooperating via the internet, whereas many open source projects are collaborative exercises between public and commercial concerns. Either way, the track record of open source software development cannot be denied and the importance of transparency in the electoral process should not be underestimated. Any legislative limitations that do exist should be overcome by a more enlightened and open approach to the regulatory system.

Scrutinised code can be verified by the use of checksums, and the integrity of the vote can be protected by secure cryptography. The real potential of free software is the precedence of peer review as it is practised in scientific research - and the simple truth which follows - that many hands make short work and "given enough eyeballs, all bugs are shallow". Peer review helps to ensure that different avenues of research are explored and that obvious bugs don't emerge. Every programmer worth his or her salt knows it sometimes helps to have someone look over your shoulder and scan your code for the obvious omission or error that you didn't see because you were too close to the code.

The report also argues that open source software offers little transparency because the visibility of the code means nothing to the untutored eye, and is an example of "security through obscurity" making the contentious assertion that "taking a proprietary software product and disclosing its full source code to the general public will result in a complete forfeiture of the software's security." If that is indeed the case then the code isn't good enough for voting purposes.

The report misunderstands the differences between patents, copyright and trade secrets and how they affect free and open source software, and makes a grand and inappropriate gesture by proposing that if the US government were to mandate the disclosure of the code used by current voting machines this "would open the possibility of a government taking in violation of the United States Constitution. The Fifth Amendment to the U.S. Constitution prohibits the taking of private property without just compensation," suggesting that disclosure of the code (and any faults in the code) that is used to register votes would be a violation of citizens' rights.

Next: The sacrifice of accuracy

« previous 1 2 3 4 next »