Mark Lemely, Doug Lichtman, and Bhaven Sampat have a sensible proposal in the winter issue of Regulation. They start with the well-known fact that most patents receive scant attention from examiners and accordingly, many bad patents issue. They then argue that this is not a problem most of the time because most bad patents are so worthless that they are never licensed or even asserted. Some bad patents do get asserted, however, costing innocent users of allegedly infringing technology millions in legal or settlement fees. They note that this is especially a problem when it comes to Internet patents, which is hard to quarrel with, I would think. Their solution to the problem is to reduce the presumption of validity for standard patents and to create a "gold-plated" class of patents that would be more rigorously examined at the applicant's expense. These patents would get a strong presumption of validity. As the authors note, the particularly attractive feature of this proposal is that it would allow those with the most knowledge about the worth of the patent to chose when to invest more resources in examining the patent so as to end up with a strong patent. As I said, an idea with a lot of merit, but read the short piece in full.
Note: Scroll down to see the article, it's the second one from the top.