Ron Brownstein, poring through the demographics, tends to agree with Ross:
If Obama runs well, he seems more likely than Clinton to assemble a big majority and trigger a Democratic sweep -- not only by attracting independents and crossover Republicans but also by increasing turnout among African-Americans and young people.
But if Obama stumbles, he could face a greater danger of fracturing the traditional Democratic coalition by losing seniors and blue-collar whites to McCain, principally on security issues. Clinton's reach across the electorate may not be as long, but her grip on her voters could be firmer.
The larger point is that each of these talented candidates continues to display stubborn electoral vulnerabilities that the other might reduce. That's why talk of their running together is likely to continue no matter how much they bruise each other in the grueling and probably inconclusive weeks ahead.
The only question I have is whether national security issues work the way they have tended to since Vietnam. We have had eight years of a classic rhetorical national security stance - and it has obviously failed in core respects. I certainly have a more chastened view of what hard power can achieve and how important soft power can be. I don't know if that has affected Americans deeply enough to change this electoral dynamic. But it is surely more of an open question than some seem to assume.