That's our thinking currently. Before we launch into a universal i-voting system, I think we need to test it with discrete groups who can be identified, can be located, and can be traced, so that we experiment to find out how well it's working.
On May 25th, 2010. See this statement in context.