It's the basic argument I've held since before the war started, but Andrew Sullivan states it nicely:

After 9/11, the cost-benefit analysis changed a little, didn't it? Who would want to be the president who gambled (in retrospect, correctly, of course) that Saddam was no WMD threat, and then discovered that some terrorist detonated a Saddam-linked chemical weapon in a major U.S. city? Do you think that president would now be popular? It's easy to know now, not so easy to have known for sure then. Scowcroft prides himself on always asking about the potential downside. Well, there wsa a pretty major potential downside of trusting Saddam Hussein in 2002. The question was never simply whether we knew the WMDs existed or not. The question was whether, without being able to know for sure, we could trust Saddam to keep such weapons away from terrorists. There's a realist case for the Iraq war: that the risks of inaction were too high, and that the threat posed by the entire region demanded a radical departure from the acquiescence to autocracy of the past.

Critics scoff now at the idea that the scraggly-looking Saddam we see shuffling into court now was ever a threat to anyone (nevermind that at the time of Gulf War I Iraq had a large, robust nuclear program, or that they experimented on rural villages with chemical weapons).