[E-voting] John Lambe's application
dglaude at gmx.net
Thu Mar 11 22:01:14 GMT 2004
A more interesting demonstration is:
Have the program behave correctly for a few minutes/hours then change
behaviour for some reason (like the first time somebody vote:
"4>11>10>9>8>7>6>5>3>2>1" and after that 10% of the vote are attributed
to candidate 4) or a timer.
It is interesting/important for the Belgian situation were sample
magnetic card are created to verify the system behaviour... then the
So if someone want to hack the system, he need to make sure the fake
software behave like the original for the fist hour. ;-)
Patrick O'Beirne wrote:
> At 11:39 11/03/2004, Michael McGrath wrote:
>> The application that records a "Yes" vote as a "No" vote and vice
>> versa demonstrates the possible flaws in the proposed system.
> Ummm, all it shows to me are two programs working as specified. It may
> surprise the completely naive, but so does the three-cup trick and in my
> opinion a submission based on such a simple demonstration would allow
> for an easy rebuttal (based on code review and test) and lose focus from
> the single issue of auditability.
> I can think of less obvious failure modes to put into the code, but even
> that is not the point. Even if the design is verified and each machine
> is individually tested to the level of the largest possible poll it can
> handle, there is still the problem of undetected operational failures.
Don't let computer expert control election...
For Belgium: http://www.poureva.be/
More information about the E-voting