I give Dragonlance a lot of grief -- deservedly so, I think -- for the role it played in forever changing both Dungeons & Dragons and the way it's been sold, but Dragonlance was merely expanding on ideas first put forward in earlier modules penned by Tracy Hickman, particularly 1983's Ravenloft. Unlike the Dragonlance modules, which, even at the time, I liked more in theory than in practice, I used to love Ravenloft. It's easy to understand why. Module I6 is a very "moody" piece of work, unlike most previous AD&D modules, which achieved their moods much more haphazardly or at least less self-consciously. Ravenloft's evocation of Gothic horror was also unlike most other modules at the time and, given my relative unfamiliarity with that genre of fiction -- I'd not yet read Dracula in 1983 -- I found it all very compelling.
There are other factors too in why my youthful self loved Ravenloft. Strahd von Zarovich, while sporting one of the most ridiculous faux Eastern European names in gaming, seems tailor-made for referees looking for a pet NPC. He's immensely powerful, well nigh indestructible, and fun to roleplay -- an angst-ridden anti-hero before White Wolf made such things a staple of the hobby. That he's the central figure in a story that provides a backdrop to the PCs' actions only made him more attractive. Moreso than most modules published before or at that time, Ravenloft is about its villain. The actions of the PCs are, in many ways, beside the point, because their sole purpose is to help to facilitate a melodrama of lust, betrayal, despair, and love beyond the grave in which NPCs are the primary actors.
And then there were the maps. Dave Sutherland's three-dimensional maps of Castle Ravenloft were amazingly innovative for the time, providing a superb sense of how all the pieces of this vast dungeon -- for dungeon is it was -- fit together. I know I drooled over these maps for many hours as a younger man and, even now, looking at them, I find it hard not to be won over by them. The problem, of course, is that, in play, they're quite unwieldy and sometimes even a little confusing. I'd go so far as to say that they're emblematic of Ravenloft itself: attractive, innovative, and a clear break from the past.
Now, I think it's all too easy to emphasize how much Ravenloft differed from its predecessors. At the same time, as I just noted, this is still, at base, a dungeon crawl and an occasionally non-sensical one at that, given that, for all its Gothic horror trappings, we find sometimes find monsters not at all in keeping with that style of writing. Likewise, there's plenty of low humor, especially puns, to be found in the module. The names on many of the tombs in the castle crypt -- "The Lady Isolde Yunk (Isolde the Incredible). Purveyor of Antiques and Imports," for example -- are outrageously bad and make Gary Gygax's own efforts seem subtle by comparison. These puns wrench one back from the Gothic atmosphere other parts of the module are trying desperately to evoke. The module uses a method of placing important NPCs and magic items based on fortune telling with a deck of playing cards. It's actually a very clever idea and, from my memory of playing the module long ago, it's effective and lends something to the atmosphere. Plus, my icy old school heart melts when random generation is involved in such a significant way.
Effective though it was, the card reading system made me wonder at the time if it was introduced partly to give the module replayability. That is, because certain important NPCs and items were placed in Castle Ravenloft randomly, the system could, in theory, ensure that each playing of Ravenloft would be different. Brilliant! The problem is that no one is going to Ravenloft more than once, because, as it is written, you can't. Dungeon crawl it may be in many ways, but there's no overlooking the fact that Ravenloft tells a story and a heavy-handed one at that. Not only does it have a prescribed conclusion, complete with Harlequin romance level dewy-eyed sentimentality, but, ultimately, what the PCs do just doesn't matter, since everything in the module is designed to support a predetermined conclusion.
Ravenloft is, like the "Desert of Desolation" series (also by Hickman -- I see a pattern here), a transitional module. There's still a great deal of old school design in its pages. There are lots of tricks and traps, for example, and Castle Ravenloft itself is a monstrous labyrinth of rooms, corridors, and crypts, making for a very non-linear portion of the game. It's also a very unforgiving module, with death around every corner, particularly if the players are foolhardy enough to try and take on Strahd without adequate preparation. Of course, unlike the later Dragonlance modules, Ravenloft can afford to be a death trap, because -- and I hate to keep harping on this -- the PCs' actions don't really matter. Strahd and his story are the main attraction here and it makes little difference whether a player loses a dozen characters along the way so long as he eventually has some character who's able to be present to witness the melodramatic conclusion the Hickmans have in store for them. That's a pretty big crime in my book and, while new and innovative at the time, it laid the foundation for much mischief later.
I still have a fondness for Ravenloft despite it all, but that fondness is born mostly out of nostalgia and that's fine. I don't think Tracy Hickman is the Devil any more than I think L. Sprague De Camp was. Nevertheless, I don't think it's possible to deny that, in both cases, these men planted seeds that would eventually bear bitter fruit. We're still wrestling with the consequences of design decisions Tracy Hickman made in 1983. The adventure path style of play, for example, is a direct descendant of modules like "The Desert of Desolation" series, Ravenloft, and Dragonlance, which represent an about-face from the more open-ended, sandbox play of the old school. The fetishizing of "super NPCs," whose actions overshadowed those of the PCs, got a nice boost too with the creation of Strahd von Zarovich. Neither of these things necessarily had to become the abominations they would one day be, but the immense popularity of Ravenloft made it hard for them to avoid this destiny. I think, with some work, Ravenloft could be remade into a perfectly acceptable and throughly old school module. That's more than can be said of the Dragonlance modules, so, in the final analysis, I'd have to say that module I6 isn't wholly without virtues, even if they are buried beneath even greater vices.