Over the last couple of days (suspiciously updateless), I've had growing concerns that I'm not quite heading in the right direction and need to re-evaluate just what I'm trying to do. The contents of yesterday's meeting (during which I admit I wasn't feeling very well—sorry, Staffan) only increased my concerns, while at the same time making me think of something that may solve it.

Okay, up to now, I've been mostly looking at this from the perspective of generating the simplest kind of quest I mentioned a couple of posts back, thus the needs for things like generating NPC needs and motivations. However, I'm beginning to wonder if this is really the most interesting path to take; actually, I'm pretty sure it isn't. A much more interesting thing to look at would be category two; “side quests”, or “branch” quests.

What drives a story? Conflict. Any story worth its salt is intrinsically about conflicts and how the characters involved in them react to the conflicts. Looking at other games, I'm pretty convinced this is also the case for games. In particular, I think they can form a good basis for generating exactly the kind of quests we're interested in.

Aside from updating the thesis document, I looked into using conflict as the source for creating quests, and I'm hopeful. Unlike the previous approach, which ran into troubles when I realized connecting NPC motivations with the quests they generate was slightly less than trivial, I can more easily conceive of a good formalization for using conflict in this way.

It should be noted that I use “conflict” in a rather abstract sense. A “conflict” here is simply two needs or facts that are at odds with each other, and not necessarily an armed conflict. Examples of such conflicts may be “Person the hero cares about needs a cure for his disease, but the cure is not available.” just as readily as “These two clans are fighting for dominance over an area”. Right now I'm considering a system functioning by choosing a “Conflict Prototype”—yes, this means I've resigned to having to involve the author in such a direct way, but in the end, I think it's unavoidable anyway—and then generating quests according to the follow rough progression:

(0. Pre-story — The things that have come to pass before the player arrives. Not actually part of the gameplay, but of the history behind the conflict.)
1. Engagement — The player is dragged into the conflict somehow. There's a wide variety of atoms that can be chosen here, such as the player being beseeched by someone in power or someone for whom the player character cares, or by being forced or coerced, or simply lured by an offer of money.
2. Proving his worth — The player has to prove his worth to the people involved in the conflict, usually just one side of the conflict. There is potential here for letting the player choose sides. This part may be small, or omitted entirely, depending on the circumstances of the conflict.
3. Aiding his allies — The player carries out tasks to improve the position of his allies in the ongoing conflict. This could mean searching for the cure to the disease, or sabotaging the operations of the opposing clan.
4. Completion — The player moves to end the conflict entirely, such as bringing the cure back to his friend or going up against the other clan's leader in a final showdown.

This particular pattern of progression seems to be ubiquitous in existing games, written by actual authors, which is why I have a good feeling about it. It also seems directly formalizable and implementable in code, which is certainly a good sign!

Will think more tomorrow.