I found an essay with so much good stuff in it that I was in danger of exceeding my quote quota, so I'm putting more quotes here:

"Awareness of what is first-best is a condition of being able to aim at second-best."

"What would be nice, then, would be a political philosophy that did a better job of taking this sort of typical deformation into intelligent account, which would discourage it – since it thrives on not being seen for what it is. (Also would be nice: a pony!) A theory of first-best that talks astutely about second-best. This is inherently hard to do, so I don’t say ‘theorizes well’. I guess I would propose a sort of line-of-sight rule. Optimally, you shouldn’t lose sight of your ideals or of reality. So much so obvious. But really the trick is keeping accurate score with regard to semi-idealistic philosophical and policy proposals. Philosophers like to talk about the difficulty deriving an ought from an is (or an is from an ought). But it’s equally important to think about the difficulty in analyzing an is-ought compound into component elements, the better to reduce and potentially reconstitute it."

"As philosophers, we would like to address the strongest arguments our opponents have. But this etiquette of the seminar room is bad ethics, in that this just isn’t the right approach in political philosophy. Liberalism is as liberalism does (not just as it might do, ideally). And the same goes for conservatism and libertarianism and communism and the rest. If you only address good arguments you will miss out on a perilously large proportion of your actual subject matter."

I'm not quite as sure about this one. I suppose it depends on what one means by an argument. Political philosophy presumably includes implied arguments that trying to approach some specific proposed ideal system will produce better results than not trying to approach it.