(Rob Conley's comment about the terseness of The Spinward Marches and how a little more detail might have improved it as a tool for sandbox play got me to thinking about my issues with most published settings.)
A cursory reading of my posts on this topic would suggest that I hate detailed settings and prefer those that are more "skeletonic" in nature. That's simply not the case, as I adore Tékumel, which can hardly be called un-detailed, let alone skeletonic. In point of fact, what I actually dislike about many published settings is not so much their detail but their tendency to add ever-increasing levels of specific detail over time, details upon which later supplements come to depend for their very existence.
Let me cite a few examples of what I mean, using the history of Traveller for my basis. When I first started playing the game, probably in 1980 or thereabouts, my campaign was set in the Spinward Marches, using Supplement 3 as its basis. The thin information it gave me was a good starting point and the fleshing out of the sector that GDW did through its adventures and supplements was slight enough that it did little violence to my own presentation of it. Then, the Fifth Frontier War happened. GDW felt that it needed to "shake up" the staid Marches by having another war break out between the Imperium and the Zhodani. The War was covered in The Journal of the Travellers' Aid Society and through a wargame, among other means.
When the war concluded, there were consequences. The map of the Spinward Marches changed and the political fallout from the War had repercussions in adventures and supplements. In short, the War changed the setting, in ways both subtle and (occasionally) profound. Now, of course, one could choose to ignore the War and its canonical outcome, but, if one did, then GDW products set in the Marches after its conclusion become increasingly less usable, or at least demand a lot more modification to suit one's own private vision of the sector. But I'm all in favor of referee modification, aren't I? Why is this a problem? Well, it's not in any absolute sense, but the more a setting expects that I hew to an evolving history that occurs independently of my own campaign's history, the less point I see in using it at all.
Sadly, GDW never seemed to learn their lesson on this score. They repeated the mistake again and again. In the late 80s, they threw the entire Imperium into a civil war (oddly called "The Rebellion") that ultimately laid low interstellar civilization first in a period of decline called "The Hard Times" and finally in an even more bleak one that more or less rebooted the entire setting. Without debating the merits of these particular decisions, what I find most objectionable about GDW's approach is that Traveller's setting changed from a wide open one whose presentation was piecemeal and open-ended to one that more or less bound you to its official development. Again, sure, you could just ignore a lot of what they were producing and use it as you saw fit, but, from my perspective, if you're going to do that, you might as well create your own setting.
To put it another way, when I started my Traveller campaign in the Marches, I could easily portray the Imperium as a distant, decadent, and thoroughly corrupt institution, like Hollywood's imagination of the late Roman Empire. The rulebook and the early supplements didn't contradict this portrayal and neither did they contradict alternate ones, making it possible for each referee to present the setting as they felt best. Over time, though, this changed and it became more and more clear that the Imperium wasn't decadent or thoroughly corrupt and was indeed better described as noble, if somewhat flawed. Likewise, the earliest presentations of the Zhodani were outright villainous and some sources even call them "barbarians," which only strengthened my late Roman interpretation of the setting. Later, though, the Zhodani were portrayed more sympathetically.
My point here is not to argue against any particular thing GDW did to the official Traveller setting. In the end, this isn't about Traveller at all, but about the tendency for game companies to develop their RPG settings in ways that both demand you buy every one of their products to stay "current" and that encourage an obsession with setting "history" that occurs independent of player action. This also isn't about the level of detail, because, as I said, I can enjoy lots of detail in certain contexts, even if it's not my preferred approach. If one looks at Hârn, you'll find an incredibly detailed setting, filled with more information and minutiae than most gamers could ever use. What you won't find, though, is an evolving history that requires one keep up with all the latest releases to be able to enjoy those published down the road. Hârn is what you might call a "steady state" setting, forever stuck in an eternal Now -- until, that is, you get the ball rolling in your campaign and nothing Columbia Games publishes will ever tell you what's going to happen after that starting point.
I guess, in the end, what bugs me about too many gaming settings is the assumption that they exist independently of being used in one's own campaign. Some likely will see this as a positive thing, since it implies a living "reality" to these places, but, for me anyway, it's a big turn-off. I don't like game companies dropping world-changing events into official settings, no matter how cool they might be in the abstract, because experiences teaches me that these events almost always make it harder for me to use future products unless I decide to go ahead and use these events in my own campaign. This isn't an old school vs. new school thing; rather, it's a situation created by the need to sell more setting supplements. I have no beef with companies wanting to sell products. I simply think there are better ways to do so without turning their official settings into novels or movies whose events are decided by someone other than me and my players.