Thad’s posting about NASA is an interesting one. If you google “space shuttle dangers” or “space shuttle risk assessment” you’ll find that NASA does, in fact, fly the space shuttle in the face of safety concerns. (They even repaired the shuttle during the most recent mission.)
The whole point is what constitutes a reasonable level of risk for the shuttle. They won’t fly if there is a measurable risk of catastrophic failure. At the same time, NASA acknowledges that it is impossible to eliminate risk–they are, after all, launching human bodies into space on top of millions of pounds of rocket fuel.
Thad’s post makes one good point: analogies to the space shuttle are silly. A failure in the space shuttle is almost immediately catastrophic. A failure on a voting machine is certainly not life threatening, and is even unlikely to turn an election.
What does constitutes “reasonable risk” in electoral systems? The worry that Thad raises, I think, is that the California Secretary of State is using an unreasonable level–no risk at all–without considering the relative risk of alternative systems.
Now, I’m no expert on the subject, and I am pretty sure that a chapter in my co-bloggers’ forthcoming book discusses precisely this issue. Maybe we can get an extract from the book in an upcoming posting.