Dwimmermount is a classic dungeon in every sense. The preview for backers of the Kickstarter has a first level that feels both formal and archetypal, yet also lavishly rich in atmosphere. There is a strange mixture of James' staid writing style and the moody ambience that quickly coalesces around the dungeon environs. His writing is effective, and creates a gloomy picture of a ruined hall, with walls and floors now too inert and mute to tell all of their stories.
As James' promises elsewhere, these rather mournful early hallways and chambers will eventually give way to something much stranger. From that discussion, and James' interest in planetary fiction, deeper levels will seemingly take a dip towards period science fiction with Burroughs et al. While it may be judging a book by its cover, from the early drafts I expect the resulting product wil be a well-woven and meticulous classical dungeon, full of ambience and character, with a few twists and turns towards the finale.
In that same discussion, James has expressed interest in using Dungeon Crawl Classics with Dwimmermount. While this is also my current game of choice, the more I read it, the less I am convinced that DCC is well suited by a megadungeon. One of the early design goals was that DCC was not D&D, and would not follow the same path that D&D took. DCC cleaves closer to pulp 1970's science fantasy literature, whereas Dungeons & Dragons became beholden to its wargaming roots. The dungeon is well designed for the later style of play, with gritty room-to-room warfare and strategic exploration and conquest of different zones.
In contrast, Dungeon Crawl Classics seems (perhaps ironically, given its name) better suited for episodic play typical of pulp novellas and the short stories found in the back of cheap science fantasy rags. DCC personas do not fight for every inch of ground with an army of henchmen, but rather they go on flashy capers, discover horrible secrets and fight gross-out final bosses.
This style is well supported by the panopoly of modules Goodman Games has lined up, but I am wondering how it would look for novice judges trying to prepare their own material. The approach I have been taking lately is similar to the method I used in an unpublished homespun roleplaying game a while back. I would build a somewhat self-contained locale (called a "setting") and come up with 2 to 6 major spots in the location (called "scenes"). Scenes were evocative locations where significant story encounters would occur (thus, extraordinary events would never happen in ordinary places, and vice versa). Different settings would be connected in a web geography according to how the narrator expected the plot to unfold (and later, redrawn according to how the plot would actually unfold). Settings themselves would be given brief and colourful descriptions, which would help the narrator improvise.
Thus, the city of Swampgut (setting) could be connected to Cairnlands (setting), which would be connected to Blagga's Hold (a setting, with the descriptors: "crumbling for centuries," "dry mud-walls," "red qwartz canyons," "ant-hole hallways" and "dry creekbed"). Once the characters beat a certain number of the scenes in Cairnlands, they could cross from Swampgut, through Cairnlands and to Blagga's Hold. The latter might have four scenes, including "last standing guard tower" (bird's eye view, crumbling steps, massive bonfire, Blagga's spy-falcons), "Blagga's Harem" (deep in the bowels of the fortress, eunuch guards, debauchery, brazen idol, Blagga's platform, poor acoustics), "black pit" (thousands of feet down, lightless warrens, lurking minotaur, sandy arena) and "master kitchen" (hundred cooks, strange meats, massive hearth, hanging pots and pans, chaos and clamor).
How the players interract with such a setting is freeform, and the judge is encouraged to briefly describe otherwise uninteresting travel. The characters could sneak around the creekbed, find a secret entrance in the wall, and creep through winding warren-like halways until they find a peephole that spies into the harem (ahem, to overhear the infamous Blagga's plans, of course). Once the players goals were sufficiently achieved (or failed), the setting would be expended and the story would advance. For settings that are pure obstacles (like the aformentioned Cairnlands), the narrator may require a certain number of scenes to be beaten to open up further settings.
This is a quick way to sketch out a basic map of the story, but it is still flexible to be modified on the fly. By focusing on dramatic moments and places, such a campaign would feel more pulpy and narrative, which can be scaled back according to the tone the judge would like to set. This approach de-emphasizes mapping, tracking of time and supplies, and the general inertia of traditional exploration. This does not provide an actual plot, however, and I hope to touch on a few tips for adventure design in the near future.