2009 makes sense, as long as they stick to features that don't require epic testing they should ship on time.  Vista was slow-going because of lots of large internal changes like entirely new display/audio/network stacks.  Probably necessary work from a long-term engineering standpoint, but hard to test, and sell.  Win 7 will probably be pretty PM-driven like Office and be all UX and fit and finish stuff.