The last UK e-voting trials, in 2004, involved electoral areas with 8 million voters and five different e-voting channels: digital TV, remote internet voting, kiosk voting within the polling station, remote internet voting, remote kiosk voting at, for example, supermarkets, and telephone voting, including voting by SMS. UK officials have also been giving advice to their counterparts in other countries for several years now, through focused events in London.

Louise Ferguson, the director of Digital Habitats Ltd, a user experience consultancy firm that looks at design and evaluation of new technologies, has studied e-voting and e-government systems in the US, the UK and elsewhere. She says governments need to pay more attention to usability and social issues when implementing new systems for public services.

"Whether e-government services or e-voting systems, the user experience is paramount. It is all about providing trustworthy systems that people find useful and usable. Lack of usability and accessibility undermines trust, and while trust is very easy to break down, it is very hard to recover," Ferguson says.

A report from the Office of the UK Deputy Prime Minister points out that the public is already happy to vote by telephone and the internet, in such TV shows such as Pop Idol and Big Brother. It also suggests that e-voting could not only improve the efficiency of vote counting but also have a profound effect on turnout, especially among younger voters.

"Personally, I regard casting a vote for a parliamentary representative for the next four to five years as a far more complex issue than voting for someone over the phone in Big Brother," Ferguson says. "There are several social aspects the must be considered. In the UK there is, for example, a tradition that you turn up in person and challenge other people about whether they are allowed to vote in that district. Many people are also entitled by law to be present to follow the election procedure, such as candidates, their agents and the press."

Ferguson says making the act of voting private instead of public could also have an impact on how people vote. For example, in a household or workplace, voters could be coerced by family members or employers.

"E-voting is not like e-banking, for example, which is pretty straightforward. A person's e-banking is all about making sure that the person is who they say they are. And both the customer and the bank can check the transaction, there's an audit trail. With e-voting, you have to have identification of the voter but on the other hand the voting has to be secret."

Recent examples from the US show that security has to be taken seriously. A report from US academics criticized SERVE, a pilot for an internet voting system, that enabled the US armed forces to vote online while serving overseas. Their report showed several risks in system security and the US government abandoned the project soon afterwards.

What then is crucial for the success of e-voting? "Well, the authorities have to address these issues rather then pretend they do not exist," Ferguson says. "There are several examples of systems, both low-tech and high-tech, that have been implemented and then proven problematic¬, so creating a backlash. We've seen this in the UK with the recent expansion of postal voting. To avoid loss of trust, system security, usability and accessibility issues have to be handled in a professional, knowledgeable manner while systems are being designed. Only once those issues have been addressed should the system be implemented. At present, it is more often a case of implement first and face the problems afterwards.

"E-voting and other technology applications are often viewed as exclusively technology issues, but it's often the people factors that make or break a system. To make a system usable, and used, the needs of the people who have to use the system are the appropriate starting point, not the requirements of the technology. To ensure good design, usability needs to be built in from the beginning, it cannot be added as an afterthought. And any e-voting system must also be vigorously user tested before deployment."

Hendrik Bergstén