A series of recent posts (Tenkar asking whether people prefer one-sheet adventures or longer modules, Billy talking about the problem of running mysteries, and Beedo's review of Peter Spahn's The Stealer of Children) led me to think about *my* ideal adventure format.
On the question of length, I'm definitely in favour of adventures that can be contained in a small-ish number of pages (probably 8-16). This isn't about the length of the adventure itself, but the number of pages over which the information is spread. As a GM I find that having the material spread over 64 pages (or more) makes it difficult to master the logic of the moving parts of the adventure; what the NPCs want, what will happen if the PCs do nothing, what will happen once they intervene, how actions in one location will affect the rest of the world, and so on. Adventures are, after all, a set of [analogue] moving parts into which the PCs intervene, and given that the advantage of a role-playing game over a CRPG or a gamebook is that the PCs are granted the freedom which means that their interventions cannot be predicted or limited, the key task of a GM is having a working model of these moving parts running in his or her head.
So, my ideal adventure would set about providing the 'logic' of the adventure in a comprehensible format designed for ready reference. Indeed, the whole adventure module should be designed as a reference guide, an instruction manual even, not as an entertaining read. The first few pages should detail the adventure locations and timeline, with the next few describing the the NPCs and organisations. Everything should be in a game-able format. This should probably be in the form of bullet points, listing key words and phrases that enable the GM to quickly grasp things such:
- Physical Description
- Characterful Phrases
As well as the likely actions / responses of the NPCs or organisations (or any other aspect of the world) to PC action. PCs do not take the optimum path through the adventure. In my ideal adventure, the writer should not encourage the GM to forbid PC action, or render player choice meaningless. So many (very good) adventures contain lines such as "make sure that the PCs spend the night in the inn. If the PCs try to leave, have X happen. If they persist, have Y happen". The world should shift in response to PC actions, but in the way that a world would, not to render PC action irrelevant.
Instead of lines such as these, by presenting the elements of the adventure as a series of moving parts, not as a path or as a series of scenes, and by presenting such information in a readily 'game-able' manner, the GM can quickly build the model of the adventure in his or her mind and by having such a model, accommodate player choice and PC action.
Of course, some of these actions will produce a 'pathetic' outcome - the PCs will fail to solve the mystery, or will accidentally-on-purpose kill the main suspect within the first few minutes, or will find some way to otherwise disconnect the moving parts of the adventure. And that is fine, as you have a world so rich in adventure that the next adventure hook is right around the corner. Don't you?
And that is the other thing that would be an integral part of my ideal adventure format, advice on what to do when the PCs fail, lose interest, chase a red herring, or suffer a Total Party Kill. The PCs should be allowed to fail, but the world should keep moving. And the way that is might keep moving, and producing adventure, should be explicitly labelled - why not have a section titled "What to do in the case of a TPK"? My ideal adventure format is more 8-16 page 'instruction manual' than 64-128 page 'fantasy encyclopedia' .
 Quite a lot of perfectly good adventures begin with several 'passages' of play in which player choice is severely restricted, all in order to get the PCs in the right situation for the adventure proper to begin. If this is necessary, play should not begin until the choices of the players are meaningful. In my experience as a player, when the opening passages of play are obvious railroads I lose confidence in the capacity of the adventure to allow me to make meaningful choices for my character/s. If it is absolutely necessary, begin in media res rather than engage in false 'play'.
 Lots of my favourite adventures are in no way presented in 'my ideal adventure format'. They are my favourites in spite of this, though, and when I prepare for a game using these adventures I find myself either working several times as hard as I ought to have to master the 'working parts', or rewriting the adventure in order to ensure that I have made the ways in which the parts fit together explicit.