I’ve been pondering relationship modeling and database elements for a few months. On the long, slow drive home this evening, it drifted further afield into the concept of a truly living dungeon ecosystem. The concept of a changing dungeon environment has been around for decades. The locale is not static — adventures and the denizens are constantly changing. So why not toss together some code and shift things about in a mega dungeon setting? Certainly not a novel idea but it would be fun to try.
Start with an open sourced set of maps and perhaps initial stocking. Then completely rethink how monsters, treasure, and such are distributed. My first cut would be encode rooms not as just a specific location but as something with volume / square footage. Groups of rooms should form zones that are easily controlled or should be utilized as a thematic region. Zones could easily span levels if the entry/exit points are contained within them. Dungeon levels are artifacts of system design. They generally contain a specific level of creature / difficulty. While that could be modeled as a concept, I believe an intelligent stocking / disruption model would serve the same goal.
Monsters details would need to grow beyond simple statistics. The region of influence they could potentially control and the desire to do so are interesting starting points. Further, other creatures they tolerate, manipulate or seek to eradicate are far more complicated factors. A group of ogres can dominate the kobolds near an entrance but are afraid of the beholder. If they have a slave or three, how do they utilize them to defend the space they control or increase their status within the dungeon?
Monsters also have metrics for desire. Perhaps they just want to control an area or grab the loot the adventurers might otherwise get. The relationships with other denizens will drive what do on a daily basis. Fear, greed, survival, etc. might all influence what they undertake. Does the desire for treasure lead them to try to overcome something else?
Treasure and traps would be constantly changing elements. If a minotaur steals treasure from a group of adventurers, he’d would control more loot enticing the greedy residents to attack him. Monsters could use traps as deterrents to enter the zone they control or as warning systems. More intelligent creatures would employ more traps but would have to do so at the expense of other activities.
When no one is around, the dungeon would continue to evolve. Based on some fixed time scale, the creatures would continue to live. They might attack others, build stuff, or undertake other activities necessary for daily life. Every action would result in changes within the dungeon which would be reflected in the current state.
An interesting concept but adventurers are the ultimate disruptive force. Why not enable game masters to enter events such as the death of creatures, stealing of loot, disabling of traps. The changes would then cause a reaction throughout the dungeon. The propagation would then be available to others or the same GM to utilize.
The base map would be unlikely to change. Everything else would need to be available as a download in PDF form. Likely in small chunks such as a couple of levels. Coding it would be damn fun. I’m dubious as to the usefulness. Probably best left as a half-cocked idea than a major effort with no consumers.