I'm studying Lars Svensson's 1997 paper on the topic right now for my final tomorrow. It's a very elegant argument. One of the nice things is that although it's a fairly stylized model you don't get the feeling he's side-stepping anything to get his result, particularly on the rational expectations front. You are targeting expectations essentially.
That of course makes me think of Krugman's work on monetary policy at the zero lower bound as well. Of course it's a little different. Svensson is treating the forecast as an instrument for targeting an optimal current inflation rate (think about it like aiming your car at a fixed point ahead of you on the road to stay in the lane, rather than frantically trying to look to either side of you to make sure you're in the lane), whereas Krugman saw the expectations as an end in themselves. But you can still see how the two fit together.
The Taylor Rule actually holds up OK because something like a Taylor Rule pops out on the other side. Friedman's approach to monetary policy does not come out looking very good at all.
Here's a link to a Cato Unbound piece that has a good section on targeting the forecast from back in 2009.