A splat book may have half a dozen really cool class ideas and ten great new feats; but if you're also getting 96 pages of other stuff you don't want is it really worth it?
Lanefan
I agree, there have been a LOT of classes published in each edition over the years, but that hasn't bothered me as much because I'm not choosing a class as often as I might be choosing a manuever or trick (or as much as I had been choosing powers). For me, quality is really at the heart of the "glut" problem (and of course, that is subjective, so it is difficult to avoid). I don't mind "splat" books and the like. In fact, I enjoyed a number of books that were produced in 3.5 because I thought they (for the most part) continued to produce new, distinctive elements to the game. 4e started out that way, then it hit a dry spell, and picked back up with things like the book about the Feywild. So quality is key. A diversity of good things is well... good, but a diversity of things - with only a few of them being good - is glut. Complicate that situation with a mandatory decision at every level (e.g. you must pick a maneuver, trick, etc...) and you've got players wading through a lot of options, but only picking a few because they are the high quality items - which only exacerbates the issue.
For me, it's not just the number of elements that are presented, but its the number of elements that provide an added, distinctive quality to the game. The question is, how can 5e provide a game experience that offers a wide variety and choice and customizable complexity and continue to publish new game material while avoiding glut?