The DrugMonkey blogger, who apparently works for NIH or another US government agency, says that the method for deciding on the order in which authors are listed on papers needs to be revised. PIs are always last, leaving everyone who did the heavy lifting vying for the coveted first-author position, according to the post. DrugMonkey's proposal: let reviewers flag which parts of the paper are most important, and then assign authorship order based on who contributed to those key parts. Judging from comments on the blog, there's quite a lot of disagreement about this "modest proposal."