Hi Phil,
Personally I think some of the problem is that there's too much emphasis on "new" work (granted, that's what PhDs are supposed to do, but still) rather than fine-tuning/collecting/merging known results into clearer, more general & powerful techniques. There is plenty of utility in having Yet Another Review article come up year 20 years or so, and it would do the average reader a lot more good than some journal article where a guy thinks he's uncovered something truly new and is documenting somewhat preliminary results in a Journal paper.