Jean Kazez makes a point I've been making about the Sam Harris and science of morality thing. I'm less mystified than Jean is about why so many atheists have lined up against Harris on this point; having argued with atheists in my time, it's clear enough that the absolute separation between morality and knowledge about facts is a pillar of many forms of atheism -- it's a common way of criticizing religion (which tends either to hold that there is a tertium quid or to deny the separation altogether). This does not appear to be generally true among atheists who have philosophy as their profession, but it is common enough outside.
Like Jean, for me the ultimate point is not that Harris is right, but that the arguments against him are for the most part puzzling at best and demonstrably question-begging or bad at best. There are exceptions (like the value-ordering arguments I've mentioned before); but they are rare. Most of the arguments are cousins of the disagreement or motivation arguments Jean mentions, which are arguments that are both very limited in what they can do and very problematic when directed against moral realism generally (rather than certain very narrow forms of it, none of which seem to be relevant here). Much of the problem is one hinted at by Jean here, which is that moral realism is strategically the wrong point at which to attack Harris. It's the point at which Harris is making claims that are very easily defensible, that are very difficult to mount good arguments against, and that will be regarded with the most sympathy by people who otherwise have nothing in common with Harris.