I met with Steve Englehart today and discussed some of the problems, and several ideas came to me. The most important is the realization that there is no general solution to the problem of adapting graphics to this technology. However, there are several ameliorating considerations. The first arises from Scott McCloud's notion of the interplay between text and graphics. He points out that text and image can share the load of storytelling equally, or one can take the bulk of the storytelling load, leaving the other free to roam. I think that this observation holds the key to our solution. I've been trying to come up with a general graphics solution that will always do the trick &emdash; instead, I need to think in terms of both text and graphics. I must use the translator code, perhaps spiffed up a bit, with the assignment that the writer must carry the story in the text. But then I also charge the artist with the task of supporting the story as much as possible. Thus, when the artist sees an opportunity to take a strong lead in the presentation, we let him run with it. But if the artist can't see any way to tackle the problem, we let him take a more indirect or abstract approach.
Another concept is graphic apposition. Instead of using just one image, we allow the artist to create as many images as he desires, and the system can randomly choose among images.
A third concept is image programmability. Here I'm not talking about something like Macromedia Director, but instead something more along the lines of a graphic analogue of the translator system. In other words, the graphic artist is allowed to specify not only the raw image, but also some other elements that are assembled to make the final image. The most obvious of these other elements are the faces of the actors &emdash; where they go on the image, what facial expressions they have, and how big they are, perhaps.
But there's more. What about borders, choice of background or primary foreground color, or use of different fonts for the textual output? Perhaps I'm tackling this backwards. The artist doesn't want to merely diddle with computer graphics options &emdash; those things can all be handled before runtime. What we want is ways for the artist to diddle with the output at runtime, in response to the situation. What factors might be put to work in affecting the imagery?
Thinking over the options, I see some possibilities for using Location as a background. (I've already mentioned this before, I think.) Time doesn't seem too helpful, unless we could key it to diurnal time, but that would show up in the background as well. No, what we really want is some kind of emotionally laden aspect to all this.
Wait a minute! I forgot POV! Clearly we want some major differences in the image depending on the point of view of the player.