Moreover, it is a simple matter to avoid state "testing" by providing via computer code that the virus would not trigger except on election day, and only when the computer has more than 50 votes on it, and is en "election" mode and not test mode. This avoids all known tests which are either not on election day, not under election conditions, or not featuring large numbers of votes like on real Election Days.
Ultimately, all such tests of voting computers are irrelevant and deceptive in that computers simply do as they are told to do, and nobody truly doubts they know HOW to count. Passing a test the day prior to the election is as irrelevant as the Tuesday bank robber proving he did not rob any bank on Monday. Rather, the question is "What were the computers TOLD to do on Election Day?" We the People are not allowed an answer to that question. And indeed, for reasons Harri Hursti can explain, we can never know for sure whether or not there was a double Trojan Horse in place, for example. See "Reflections on Trusting Trust." http://www.acm.org/classics/sep95/
THE COUNTER-CHALLENGE
In light of the above important context, a very realistic security test would be to allow any citizen to hire the computer expert of their choice, and give that expert days or perhaps weeks to hack the system, BEFORE any seals and so forth are placed on them.
This is just like the real world in which top elections officials have virtually unlimited access to the machines.
Next Page 1 | 2 | 3 | 4 | 5 | 6
(Note: You can view every article as one long page if you sign up as an Advocate Member, or higher).
