Kevin Drum, citing 1968, says that Democrats needn't panic. I don't really buy the historical analogy for a reason that can be summed up in two words: George Wallace. Without him in the race to split the white supremacist vote that went for Goldwater in 1964, Nixon almost certainly wins in a massive landslide. For this reason, I don't think that this is a very encouraging precedent. On his overall argument, however, I agree with him at least to a point. I don't think that the extension of the campaign per se is a big deal at all. Were Obama to win Pennsylvania and end the race, for example, I don't think that the extra month of campaigning would hurt him much as a candidate, and as Kevin says depriving McCain of oxygen may even be a net positive.
The bigger problem is a scenario (which, given that Clinton has to be considered a strong favorite in Pennsylvania, is a substantial likelihood) where Obama has a lead of pledged delegates in the high double or low triple digits but is coming off some high-profile losses in state popular votes. Serious attempts by Clinton to seat delegates based on the results of Michigan and Florida straw polls (although not necessarily a re-vote agreed to by both campaigns and the DNC), for example, would produce very serious conflict. And if Clinton were put over the top by superdelegates (which I continue to think is highly unlikely) she would be severely weakened as a general election candidate. Is it possible that the Democrats could win after a protracted convention battle? Sure. But it would turn an election in which the Democratic candidate has considerable structural advantages into a much dicier proposition. I don't think that this is something to be sanguine about.
But that's life; politics is messy. Having two very strong potential candidates has its downside. It's especially ironic that a process re-designed to produce a quick victory and party unity has led to the opposite, but unintended consequences are endemic to political institutions.
--Scott Lemieux