What Story? Reporting a MUD-Dev Thread from April – May 2000 [re-post]

by wsampson

This is a re-post from my original on the Preserving Games blog, January 23, 2009. I expect to re-post most of these from over time. Looking at this one, I’m struck by how the posters’ concerns are still being considered and played out. In particular, I’m interested in the incidental storytelling notion, in my experience best demonstrated by Dwarf Fortress. A good example of that emergent storytelling is nomad (tim)‘s great illustration of one particular battle.

From August 22, 1996 to the early October of 2004 the MUD-Dev mailing list housed a slew of earnest and lengthy discussions — technical, philosophical, design and otherwise — concerning the development and play of multi-user dungeons (MUDs).

MUDs were developed before graphics-capable computers, originating and remaining entirely text-based. Players type commands, role-play or just socialize through the keyboard. Gameplay ranges from hack and slash to interactive fiction to simple chat, with all sorts of degrees of role-play in between. MUDs are often discussed as the precursors of today’s MMORPGs, and therefore an early index of the concerns and goals of those MMORPGs by at least 10 years. But MUDs are by no means a dead genre. New MUDs, and players for them, are continually coming into play.

The MUD-Dev archive is available at Raph Koster’s site as a bundle of distinct HTMLs, or right here at the School of Information on a single (very) large page with the messages hyperlinked and separated.

I’ve taken a look at the threads “Procedural Storytelling” and “A footnote to Procedural Storytelling”, which had sprung off from an earlier discussion, “Self-Sufficient Worlds”. The topic, broadly, concerns the art of storytelling in MUDs and more typically, how to generate them automatically on a large scale by abstracting them. Instead of doing a play-by-play of the discussion, I thought a report of the salient ideas and contentions would be more digestible.

  • The first idea to be discussed is whether, in the design of MUDs and their story content, “interesting stories” can be computationally generated. It’s posited that interesting stories are not computable and no amount of computation will generate them, at least not consistently enough for actual use. The counter argument believes that it is simply a matter of computational scale.
  • The issue of computational power and its potential for procedural storytelling is complemented a argument: automation and generation is less a problem than a toolset that allows players to interact with a persistent and consequential environment. This idea is revisited a few times throughout the thread. Posters wonder how the history of a world could be more thoroughly fleshed out, if a model could be made to allow future events to reference or be contingent upon past events, and the consequences of a persistent world given the unpredictable nature and large number of player actions any MUD would contain. I think this is an interesting consideration; it well highlights how ahistorical gamers’ interactions are in even very complex worlds.
  • The thread also concerns stories from the player’s perspective. It’s argued that a story-server would not need to generate stories per se, only conditions for stories. It would not need an explicit story model because it would create relationships that would allow a story to ensue incidentally for the player.
  • This relates closely to another idea in the thread: a priori vs. post facto storytelling for the player. A poster reasons that post facto storytelling is typical for players in most MUDs: they have a gaming session, and construct a satisfying narrative for it after the fact. A priori storytelling, however, is the conceptual model of storytelling for both players and designers alike, where a story is pre-designed and recognized as a story before it begins.
  • A related idea is player recognition of stories. It’s argued that stories rely on a great deal of context for their proper recognition and meaning, and that communicating or creating this context would be the most difficult part of serving stories to users. Tied to this issue of recognition as well is another poster’s argument that stories derive their interestingness and meaning from an emotional connection to the audience. MUDs generally invoke reward-satisfaction and excitement/adrenaline rushes through mortal-danger situations. Since this is the emotional breadth of MUDs, and since this is acceptable and expected by players, storytelling in MUDs will be very limited and difficult to introduce. It also explains the scale of drama common to MUDs: ‘great evil’, ‘save the world’, ‘escape from death’, etc. This is the most common type of drama accepted.
  • Another point is the nature of goals in MUDs. It’s argued that greater storytelling will be achieved by creating goals that are not strictly individualistic. Players belong to different groups or locales, and these entities are affected by other groups and locales. There should be a consequential dialogue between a single player’s goals and his or her community; this allows for a web of individual-group tension as well group-to-group tensions. Players therefore must consider their group(s) and others’ groups in their own actions. I really enjoyed this discussion too. Game designers are still exploring this area, allowing the player less autonomy and loading their actions with greater meaning for the groups surrounding them, or for which they are a part.
  • A tangential subject was the expanding computational abilities of computers, and how much that phenomenon should or can be relied upon when attempting to improve gameplay (in this case the storytelling aspect). This recalls Stephen’s post concerning how designers create within or against these limitations. The argument is made that the computational abilities of computers has been a continually expanding horizon, and that this allows designers to repeatedly identify computational might as the fundamental mechanic by which games can become better, as opposed to focusing on problems or innovations that do not require any more computational power than is presently at hand. Again, a wonderful point. Designers may lean heavily on Moore’s Law.

This peek at the MUD-Dev archive gives one an idea about the mental models, assumptions, aspirations that both designers and players bring to MUDs and games in general. It would of course be excellent to hear from those involved in game creation about these concerns (and terms) in relation to their specific game, but this thread archive clearly demonstrates how involved and thoughtful players were (and are) about their games.