When people say "X is possible", they mean "X has a non-negligible chance of happening in practice", and is therefore worth worrying about.
They do NOT mean "Prob(X) > 0". You're already on the wrong track if you're talking about mathematical proofs of "possibility" in engineering.
It's really, really, really small, with millions of successful flights done without problems. Yet accidents like those have happened, and have been prevented after that rule was put in place.
I was using mathematical proofs as an example of why in IT security, often times since you can't be sure, you do the proactive thing and don't run services exposed to the internet. If you have a database with credit card numbers, you don't let the DB listen for connections on a public IP, not because you think there is a security hole in the DB, but because you can't be sure there isn't.
Links please?
The rule you point to was put in place because evidence strongly supported the fact that distracted pilots are a hazard.
No such evidence (that I know of) exists for portable electronic devices, and I don't believe for a single second that Boeing would make an airplane susceptible to harm by cell phones.
If you want links to examples of PED interference, I've posted them elsewhere on this thread. All anecdotal of course, but there are enough examples that to say there's no chance if interference would imply a secret conspiracy amongst pilots to file false reports.
And again, you are misunderstanding the concern. It's not whether a cell phone can stop the airplane from working and make it fall out if the sky. It's whether, at a critical time such as during landing and takeoff, it may cause interference that can increase the pilots workload and contribute to an accident.