Dan Balz has a piece in the Washington Post about how, though he thinks that the present "political climate" seems to favor Democrats in the abstract, there's still some room to doubt if a particular Democrat wins election to the presidency over a particular Republican.
Well, I jumped into the Wayback Machine and found a poll from September 1999 (closer to the 2000 presidential election than we presently are to the 2008 election):

If Vice President Al Gore were the Democratic Party's candidate and Texas Gov. George W. Bush were the Republican Party's candidate, who would you be more likely to vote for:

Bush	56%
Gore	39
Sampling error: +/-2.5% pts

And we all know that's how the results came in on election day! And as late as December 2003, Howard Dean led John Kerry (and a host of other Democratic hopefuls) 23%-4%.
All this is not to say that a Democrat will win in 2008 or that a Republican will win; nor do I wish to say that political prognostication is worthless (after all, prognostication shapes how we take actions now, which will affect political conditions down the road). But there is a lot of time left between today and the primaries--and the general election. And a lot can happen in that time. So it seems to me that neither side should get too comfortable...