Islame,
"It would help if you could answer a question for me - what percentage of todays muslims do you see CURRENTLY as a threat to your life if they were in your vicinity (I accept that you believe that your real fear is that they could all turn that way or we are unable to distinguish between them)"
Your parenthetical part answers your question. Since I have no way of knowing which is which, I rationally assume all are a threat.
Ok, this is awesome. Let's take it another step: over here we had this thing known as the Port Arthur massacre. This involved a young, white, blonde surfer going nuts with automatic weapons. Obviously some young, white, blonde surfers are dangerous.
Since I have no way of telling in advance which particular young, white, blonde surfers are dangerous it would be "rational" for me to assume (by your standards, of course) that all young, white, blonde surfers are dangerous.
They should obviously be denied all access to firearms, and should preferably be deported. The fact that the majority of them do not appear to condone massacres committed with automatic weapons is meaningless, for the sneaky critters may well be engaging in taqqiya. You can't tell.
This is a statistical point whose principle applies in all situations where some factor, amid a general phenomenon, poses a threat of some kind. In terms of this principle, we can divide instances of this roughly into two categories:
1) where it would be irrational to apply the broad brush
2) where it is arguably rational to apply the broad brush.
In #1, an example would be the dangers of air travel, where planes have malfunctioned and crashed. The person who has a phobia about flying knows that most planes will not crash, but that doesn't assuage his fear that his plane might be the one to crash -- so he applies the broad brush approach to all planes.
The crucial question here is what factors arguably warrant treating a particular phenomenon with the broad brush approach, even though that phenomenon superficially resembles the cases where statistically the threat potential is not sufficiently high enough to warrant such a broad brush approach? With the phenomenon of the problem of Muslims, the ostensible fact is that out of approximately 1.2 billion Muslims, hundreds of millions of them are not doing anything directly dangerous to non-Muslims. If a person were to stop there and not research the phenomenon further, and simply conclude that the threat is exactly as minuscule as it appears on the surface, he would be behaving remarkably irrationally in my estimation after having researched the phenomenon further myself. This is a complex issue. For now, I will just say that the assessment of the threat of a phenomenon has to factor in qualitative factors, not just superficially quantitative factors.
Ok, but one of the qualitative factors that you have to consider, and which you seem remarkably averse to considering, is that the majority of Muslims have no interest in becoming terrorists. I agree that this is deplorably lax of them and they should be more enthusiastic, but what can you do?
Let's take your example of air travel. You say it is irrational to fear all air travel because the majority of flights don't end in disaster. You then turn around and claim that all Muslims must be regarded with suspicion and deported, despite the fact that the vast majority of interactions with Muslims don't end in disaster. Why is the former irrational and the latter rational, in your opinion?