[E-voting] Minimum requirement to juge voting "technology"

[Colm MacCarthaigh]

On Fri, Sep 10, 2004 at 02:06:00PM +0100, Timothy Murphy wrote:
> > Fraud/Bug/Fault can take place without any doubt.
> 
> Nobody is going to risk several years in prison
> in order to fiddle a vote in such a way that it may make no difference.
> 
> There are ways to cheat with the present system,
> eg voting in the names of dead or absent voters,
> but no-one thinks this justifies random re-counts.

Random re-counts would make zero difference to those threats, so
please - try not to be ludicrous.

In a traditional count, everyone can see the counting, we can observe
with our own eyes how close a call might be. If a result is within
a few dozen votes, it's very reasonable to suggest it may have been
miscounted, and to ask for a recount.

With a computer doing the counting, noone sees it. We just get an
end number. Opinion polls are very frequently wrong, or too close
to indicate, so how do we know if it's reasonable or not to suggest
tampering has occured?

I mean, if the computer says candidate A won by 1,000 votes, that's
no different than if it says they won by just 1 vote. How do we decide
which is a reasonable circumstance?

Under the current system, a returning officer can go "There's no way my
staff counted 1,000 votes wrong, as you can only have seen for
yourself.". With electronic counting, what is there for the returning
officer to go on only blind trust in a computer program?

> You can't lay down rules like that.
> In the end you just have to say that if a reasonable person 
> has reasonable doubts then the matter should be looked into.

What constitutes a reasonable doubt? Gut feeling? Prior opinion
polls? What?

-- 
Colm MacCárthaigh                        Public Key: colm+pgp at stdlib.net