This was originally a rec.arts.int-fiction post in response to the question “where do you get ideas for IF?”
I take a notebook with me most of the time and use it for working up ideas. Usually it also contains some non-IF notes as well (academic work, other projects, what I need to buy for dinner, that sort of thing). That way if I happen to see something interesting at the bookstore or library that gives me an idea, I can make a note. Once I’ve started working on something, I keep thinking about it all the time in the back of my head, and whenever I run across a setting detail, bit of imagery, coding concept, or funny-looking character on the bus that seems to fit into the project, I note it down. I remain in this mode until the game is in beta-testing: you never know when you’re going to need another touch for the setting here or there, so I don’t assume I’m done just because I’ve started coding.
But those are ideas in general, not the idea at the core of a game. Usually I start from some concept about what the player’s experience should be like, such as, “the player will talk to NPCs, asking questions and gathering clues until she solves a mystery; the conversation will work as a combination of ask/tell and menus in order to make investigation possible”.
This determines how the world will be modeled, what kind of story and puzzles (if any) naturally arise from that, and how the bits fit together. All of the setting details, plot events, puzzles, etc., are incidental, and can be discarded and replaced if they turn out not to be suitable, but the central idea has to be sound. In static fiction you might start out by identifying your central conflict and the shape of the plot, but with IF I find it’s hard to develop the plot fully before I’ve come up with the model of interaction.
That doesn’t mean, of course, that the plot doesn’t matter, or that you shouldn’t have a good one, or anything like that; but in my experience, at least, it’s fruitless to come up with a great conflict and resolution if I have no clue how to involve the player in it. At best I wind up with IF that’s an okay static story that happens to have prompts interspersed in it; at worst, the results are totally unplayable. Usually I start writing or coding, then hit a point where the project has become hopelessly muddled because I’m trying to serve two masters at once — the interaction and an unrelated plot — and the game feels boring and sluggish because it’s unfocused. Then I get sick of the whole thing and quit working on it I have far more dead projects on my hard disk than I have released games, and I’d say the majority of them are victims of insufficient clarity at the concept phase. Your mileage may vary, of course — some people apparently do successfully approach IF from that direction.
With the interaction concept in hand, I rough out how I’m going to code it, and also come up with plot events that work with the interaction. I think that has actually gotten harder for me, not easier, as I’ve written IF. I’m not sure whether this is because I’ve chosen more ambitious projects, because I’m losing my grip, because I ran through all the workable ideas I had and went on to the unworkable ones, or some combination of the above.
Anyway, once I’ve got all that, the rest is comparatively easy. Savoir-Faire was probably the most precisely planned of my IF, in the sense that I came up with the concept, made a list of all the puzzle ideas I could think of, arranged them loosely according to complexity so that the pacing suited me, and then mapped out all the puzzles in a sort of flowchart. And then I went through and coded them all until I was done. Even so there were some things I threw out in the eleventh hour, but it was a fairly directed experience.
Anything I do with conversation tends to be more organic: I may have a general plot direction in mind and some topics for discussion, but the content of the conversation itself seems to develop during coding.