Research

Research is an obvious and important step for a historically-based game. (What kind of boots did a well-brought-up young lady wear in Victorian England?) Perhaps less obvious is its applicability in developing a fantasy setting.

When I'm in the process of putting together material for a game, I sometimes spend a fair amount of time in bookstores and libraries, paging through anything that looks relevant. For Metamorphoses, I read up on Renaissance philosophy and the history of science. For other games, I've looked up architectural forms, literature, mythology... anything that seemed relevant or related in general type to the setting I was working on.

Sometimes I'm working on background and I'm trying to remind myself of the kinds of questions I should be asking. It's easy to write a room description that hasn't been thought out enough; it's easy not to do enough of the work of imagining. Looking at a book about, say, everyday life in ancient China makes me aware again of some of the relevant questions: what did the people here eat? what did they wear? what were the politics? Obviously the depth and complexity of the background you're building will determine what questions are useful ones to ask yourself.

And sometimes I don't even know what I'm looking for, exactly. Sometimes I just page through things: travel magazines; atlases; historical documents; illustrated books of antiques. Encyclopedia articles on costuming through the centuries. Web pages on 19th century railways, or out-of-fashion cars, or demonology, or the elements. I think of it as a kind of seeding process: random number generators need somewhere to start, and so does the image-generating part of my brain. Often some strange landscape will catch my eye, or some building, and even if that never makes it all the way into the game itself, there is a quality of it that informs my sense of what the feel of the game is about. For one of my works in progress I keep a collection of photographs of various objects that I consider in some way relevant: things that I imagine could exist in the game world even though they may not actually show up.

Sound like a lot of work? Perhaps it is, but it's largely fun work -- and, at least for me, corresponds with the kind of thing I tend to do by way of procrastination and relaxation. I like going in bookstores and wandering around pulling things off shelves, letting my imagination wander. It's not a lot of additional effort to take the interesting and curious things I run into that way and see whether they fit in anywhere with my work in progress.

Constructing Your Map

Structure

I like a game where you don't have to make a map.

Partly this is because I am stubborn and partly it's because I'm lazy. And I'm not sure that it's a principle to which all IF authors need to cater, or ought to cater. But there is an underlying idea here that is good, which is this: in the ideal IF setting, the parts of the setting relate to each other in comprehensible ways. Things are located sensibly. I dislike mazes not only because you do have to map them but also because they interfere with and scramble up the intuitive sense of place that I otherwise build up as I play.

The more fantastical the setting is, the easier it is to get away with making it illogical: a game set in an old western town needs its saloon, its bank to be robbed, its dusty street, and the relations of one place to another are probably clear. But a game set in some surreal landscape can conceptually get away with having, if necessary, a desolate waste full of ominous mirages right next to a beautifully trimmed hedge garden.

Nonetheless, it helps the player if you don't give him too huge a landscape to wander through at once: too much freedom is overwhelming, and one is left without a sense of where to go. Curtailing someone's motion through the landscape, at least on the first runthrough, is often useful. The idea of locked doors and keys may be a somewhat overused one, but there lurks an important issue there: in order to control the pacing of the game and give the player a comprehensible experience, you need to dole out the pieces at a reasonable pace. Not too much at a time.

Correspondingly, the more detailed the environment is, the bigger it will feel, in terms of exploration-potential. Five rooms containing one item each will seem from a game-play point of view like less work than one room with several dozen items.

Another useful technique is to create obvious divisions or symmetries in the game world, something to give the player a handle on distinct sets of locations. Metamorphoses has a very symmetrical map, because it suited the symbolism of the world to make it that way. One of my other works in progress is less symmetrical, but it has sections, what I think of as neighborhoods; each has distinct thematic features of its own.