I was recently involved in an online discussion (unexpected and bizarre virtual slapfight is a better description) about the role of a gamemaster (GM) in a roleplaying game. The discussion sparked from a comment made about Tracy Hickman and the old Dragonlance modules, which are notoriously "railroady" modules, i.e., if played as written, they are really only allowed to unfold in a single way. Many (if not most) players dislike playing in such an environment, because being told that strategies they might come across arbitrarily fail because they're not the "golden" ones that the author and/or the GM anticipates and "writes" ahead of time is not very fun. I've been on the recieving end of a railroad at least once that really comes to mind, and it kinda sucks.
Anyway, I made an offhand comment following a description by one guy that those modules sucked and that it was the module's fault. I firmly believe that it was the GM's fault, and here's the reasons why.
When someone takes on the role of the GM, they also take on the responsibility voiced by Harry Truman: "The buck stops here." The GM is always responsible for what happens at the table. He can't blame a lousy module, because one of his prime responsibilities if presenting a good game to the players. This means, at a minimum, adapting the play experience to what the players want, but ideally it means screening and modifying the plans for the session. A module that's clearly a railroad with a group that doesn't like railroads (i.e., most groups) needs to have the railroady aspects jettisoned and the GM needs to allow other methods of reaching the conclusion of the module.
My own philosophy, not that I'm suggesting that this is the "best" merely what works best for me, is to read modules and adventures with an eye towards taking vignettes; interesting locations, encounters, NPCs and concepts, but recombining and "kitbashing" them together into my own creation that best suits the needs of my campaign.