From what I’ve seen, I think Chaim’s work is fantastic, a rare example of someone actually writing new software to realize new ideas in non-trivially interactive, visual, computer-mediated narrative.

Chaim, or Michael, how generative would you say Comic Book Dollhouse is? Or, where/how would you imagine making it more generative?

over generation. That was a conscious decision on my part, since many other

systems emphasize generation over authorship and/or user experience.

I actually haven’t played around with the generative part enough to say how

powerful it is. It’s certainly a simple system, and easy enough to author

things in (though the author- interface for it isn’t very polished). The

system examines the history of the story for plot patterns it knows, and

suggests the remainder of partially matched plot fragments to players.

During my defense Will Wright wondered what the optimal ratio of characters

and props to plot fragments is, and said he was building things in CBD to

get a feel for the system’s balance. As a game designer or storyworld author,

a key concern is the ratio between author time invested and experience payoff

for the player.

Generative power is one way of thinking about the author time to player time

ratio. Another way of thinking about this ratio is in terms of how easy it

is for players to assemble stories from characters, props, and stickers created

by authors. That was a primary design goal for CBD, and it seems to be very

good at that.

I’m sure there are lots of ways to add generative power to CBD, but the questions

I’d ask of the features are:

- how much do you improve the author time to player time ratio?

- how much harder do you make authorship?