That's some interesting discussion (although I didn't read the OP's quoted letter as it seemed pretty racist).
I agree that some of the tenets of Islam are problematic, converting by sword being the main one (everything else can probably be sorted out over time).
But (to take an example closer to home), we can see it's perfectly possible to be a fundamentalist Christian tea-bagger and still not turn the other cheek, or support universal healthcare (things that Jesus was a fan of). Moreover, religiosity in the Western world has been going down on average for some time.
So it is possible for people to become less religious, or at least act less religious. What we need to do is reduce the average religiosity of the Middle East. Please try and think of this on a population level, rather than an individual level. Some people can get more Islamic, as long as more become less Islamic.
I think that what is often ignored is the environment in which Islam exists. You have a largely uneducated populace with very little social mobility (largely thanks to the West pumping ridiculous amounts of money into the ruling class who fritter it away in non-functional exorbitant displays of wealth; see:
http://en.wikipedia.org/wiki/Burj_Khalifa).
Now if there's one thing that makes people put their faith in religion, it's being ill-educated. So lets educate them, right? I'm not even talking about going in saying "this is why Islam is wrong", just giving them the basic education that we access to, and maybe throwing in there that a secular life is possible. Is that practical? Maybe, it would probably take concerted international pressure over a long period to achieve though, and everyone's a bit too bent over the barrel for that kind of thing
Another factor that we might look at (putting on my psychologist hat) is the effect of fear. If it's not apprent to everyone that it's all but impossible to beat the religion out of someone, here are the theoretical underpinnings.
The effect of fear is predicted by the reasonably well-supported terror management theory (TMT; see:
http://en.wikipedia.org/wiki/Terror_management_theory). Essentially this theory asserts that religion is a worldview meant to alleviate our fear of death, because we go to heaven/get reincarnated/whatever, so we don't actually die. TMT also predicts (and this has been empirically supported) is that the more imminent death is (higher mortality salience, technically), the more religious we get. Now imagine the effect on mortality salience of near constant war on religiosity. Now imagine the effect of the uncertainty about whether an armed drone is hovering beyond sight in the clear blue sky about to reduce you to a gurgling mess of entrails.
Not perfect, by any means, but these are some ideas for helping with the whole extremist problem. Unfortunately these solutions would take a while to have an effect, and probably won't have people waving little American flags in one election cycle (much less advance the agenda of the military-industrial complex). Easier to just say "duh doi, let's kill a quarter of the world's population"
