I was thinking about settings the other day, and I started pondering the real difference between settings based on existing fiction versus setting designed for gaming from the ground up. Some of it is obvious: Game settings are built with a certain amount of expected expandability, if only because more books can be written. They are expected to have the kind of fractal depth of history, with only the success of publishing deciding how deep into the fractal things go.
Settings based on fiction are built differently. They have an essential, recognizable seed which needs to exist to be considered part of the setting in question. In popular settings (Star Wars and Dragonlance spring to mind), the material may go quite far afield from that core, but those variations are usually just shadows of the core. The odd fan might like them better, and more details is always appreciated by a certain type of enthusiast, but the core remains recognizable.
All this was going through my mind when I got to thinking about Amber. See, something that you will notice very quickly in any Amber RPG community is that it's all about that group's version of Amber. This is not accidental - most groups are well aware of other campaigns and ideas out there, ideas contradictory to their own - but there is very little sense that other people's interpretations in any way lessen your own. When you go to a convention, the number of interpretations you will be exposed to can be truly dizzying, but what's perhaps crazier still is that they will almost all be recognizably Amber.
This resonates on some level with the classic "shared" adventures of D&D's past, the ones that the 3e series tried to recapture the spirit of. Against the Giants, Slave Lords, Temple of Elemental Evil and others were seminal experiences for players, but the expectation was that other groups played these adventures too, and had different experiences with them. Even Dragonlance, which has an "official" outcome, can create an experience through play.
What's weird is that I find this attitude to be more of an exception than the rule in games with strong settings. The idea of the official canon is such that variations on setting are pushed to the fringe, sometimes treated as a sign of an inability to "do it right".
My suspicion is that a lot of it is a function of "living" settings - people are not necessarily comfortable diverging from the game as written until the final input is received, and since many games are infinitely open ended in their presentation, that final input never comes. Some of it is also a function of volume: It is easy to retool Amber because there are relatively few moving parts. Similarly, an adventure makes for a very small setting.
I dwell on this because I really, really like the Amber approach. I love enthusiastic setting ownership. I love to see what people do with a setting to make it sing as much as I do seeing what people do to make rules work better for their table. The question, and this is the one I'll be thinking about for a while, is how to make it happen.