Barack Obama reaps the harvest of his campaign's idiotic decision to start releasing oppo research on Paul Krugman as the latter unloads on Obama, slamming him as "the anti-change candidate" who's such a prisoner of his desire for good press coverage that he's ignoring a vast populist tide sweeping the country.

Krugman's spot-on in his argument that what the Des Moines Register sees as the problem with John Edwards is, in fact, what's good about Edwards. But I think he's let his taste for revenge (understandable! I'd be really pissed if I were in his position, too) undermine his perspective on the objective realities. As John Edwards himself has said the most dramatic contrast between his vision of sweeping change comes from Hillary Clinton, not Obama. As Krugman is usually at pains to point out, there's more to life than campaign rhetoric; people have records and so forth that can be subjected to scrutiny. Nobody would appoint Mark Penn to run their political team if what they really wanted to do was lead a bold populist revival. Similarly, it's objectively true that the next president's ability to bring about big-picture change in American domestic policy will be limited by his or her ability to secure Republican votes for his or her agenda. I wish that this wasn't the way that American political institutions work, but it is. I like Edwards' rhetoric about taking down a corrupt power structure a lot more than I like Obama's kumbaya talk, but any president will face the same institutional set-up and the real limits it imposes.

All of which is to re-enforce what I said previously about voting tactically -- I share a lot of these concerns about Obama, while over time Edwards has dispelled most of my concerns about his foreign policy (and it's on these issues where the candidates views seem to contrast most sharply) -- but I think either would be clearly preferable to Clinton from a progressive point of view and the Obama campaign's poor handling of its relationship with the country's highest-profile liberal columnist shouldn't obscure that.