This measure is troubling in a few ways. One, it threatens to prolong the crisis by hiding mortgage problems in the government's books. As I've mentioned before, the danger is the potential for a Japan-style prolonged recession. This plan saves the worst offenders, while doing little to force banks to deal with their losses and move on. Moral hazard remains a big issue.
Two, it's not the best use of taxpayer money. Yes, this deal could pay for itself over time, but there are better ways to command the government's influence properly. Zingales and Krugman are both against the deal and have some interesting alternatives. Krugman would rather the government act as Sweden and nationalize firms to float them at a later date--treating the government as sort of a private equity buyout firm of last resort. Zingales would use the government to coordinate moves such as eliminating dividends, raising equity, and debt restructuring.
The nice aspect of these deals is that they leverage some of the strongest capacities of government: Coordination, transparency, and liquidity. Friedman and company tend to categorically despise all government action (except of course on financial intervention), but there are clearly many different sorts of intervention. Personally, I don't care as much about the raw amount of government spending as the form that government regulation assumes. Paulson's plan calls for a massive government bailout without holding firms accountable. Other plans envision a light government footprint paired with reasonable interventions playing on government's strengths to force firms to solve their problems themselves.