Rick Smith's Report

by Rick Smith

This was my first trip to Phrontiserion so I cannot compare it to last year’s event but I’m planning on coming next year! I am a professional game designer and my experience with IF are some of the story games, and a number of hypertext novels (which I disliked). The story games are dying out in the game industry and so I feel that the field needs a good shaking out.

I will write up a short summary of the things that were interesting to me. I’m not making any attempt to cover the events of the whole weekend accurately. This is largely a list of what I learned or cool new mental models for thinking about Interactive Fiction (IF) that I came away with.

After introductions, the weekend started out with a discussion of different aspects of IF. The first point that struck home was that a lot of people were trying to write an interactive fiction novel, where we didn’t have a good IF short story, or even a haiku. I partly disagree, feeling that various computer games might fall into that category, but there is some truth behind that point. Novels, computer games and movies have made audiences accustomed to entertainment that lasts 2 or more hours, while this may be too big / expensive for the early Interactive Fictions .

Audiences expect quite high production values which means real money. I suspect that amateur efforts for smaller budgets will have to explore the possible space of IF. Only when we have more experience with what works will the field gain larger audiences.

I believe it was Tom Gillespie who said that a novel was a poor mental model for IF. He suggested Jazz, where the musicians add to each other’s musical explorations. I think this is a very good point, but how does something like that work? Storytelling has been an author speaking to an audience for a long time, and my mind keeps falling into that world view.

We discussed a game called "Bad Day on the Midway", where you possess various characters. The interesting thing about this game is the random thoughts of those characters which appear and fade away in the screen, which gives the player a very strong sense of being inside their head. I don’t feel that this is the best model for IF. Being a (largely) passive observer with a wandering viewpoint while the story drifts along is not what I want to experience as an audience member. I prefer the computer game model where I CAN DO things!

Laura Mixon - Gould (a professional writer) made several points that stuck in my mind. She said that IF (like literature) needs to have a plot and an underlying theme that give depth to what is going on at the surface.The trick is to keep the IF player engaged in a situation of escalating tension which is the key to a traditional plot. There was a lot more good stuff on literature. At the time I was thinking something like, "Yes that is nice. But how do we make it work interactively." I kept coming up with ways to improve a story type computer game a bit, but nothing earth shaking.

Chris mentioned that the data models for traditional media are well suited (adapted) to the media. For example, the computer data structures for film (one picture after another along with sound) is well understood. Movies have adapted to (and adapted) the technology for telling linear stories in that media. This has not happened yet with IF.

The usual model for IF is a story tree, where each significant plot point will lead to a different story branch. As Chris pointed out in the Journal of Computer Game Design, this leads to the size of story trees geometrically increasing as the story length increases. I brought up the very popular Japanese comic book games, that for all their strengths end up with very sparse trees. (This type of computer game in non-existent in the USA / Canada.)

Chris suggests that the data structure for IF should be something based on what dramatically drives the characters in the fiction. His technology is based on models of the character’s personality and a simplified language that they use!

A fair bit of time was spent discussing the idioms that could be used for basing IF on. Traditional story games had ’adventure areas with gates’, some games used the ’possess characters’ model, Live action Role Playing Games (with Palm Pilot tech so that the GM’s can interact in real time!) were suggested. The ’mystery dinner train ride’ and ’host your own mystery’ games were suggested as a mental tool. Someone suggested that traditional character sheets in a role playing game were a contract with the game master over what sorts of behavior were allowed. In multiplayer IF, this sort of contract would seem to be required. Massively multiplayer role playing games (such as Ultima Online) might be a model but the objection was raised that they don’t have a plot. Ah, but they can have story arc’s that affect dozens or hundreds of characters, and we may find the first big IF are formed in a linking web of these story arcs (or at least so I argued).

An idea was floated. What about ’Titanic’? There is a major plot that can’t be affected, but characters affect dramas can take place within that plot arc. We all chewed on that for a while.

Bob Bates had some nice points that suggested that Voice Recognition would revive the text based story games. Chris worried that the ’parser problem’ (see Understanding Interactivity) would make this difficult. Later Bob showed that a reverse parser (that visually prompted the user with what words could now legally be spoken) would get around Chris’ objections. I suspect Bob is right and look forward to some of these games.

Some time around then Bob pointed out a frustrating problem I had with the game, "Monty Python and the Search for the Holy Grail" was poor design on their part. (A discussion of the damn killer bunny.) His suggestions on how to fix it, (multiple ways to solve problem which open up as the time the player is stuck at one point increases) was a really nice summary of how it should be.

In the afternoon of the second day Chris and Doug Sharp were going to discuss their story building technologies. However, Doug had gotten sick, and so Chris spent the whole time talking about his Erasmatronics. Generally the talk was oriented to ’artists’ without programming backgrounds but a little bit of Chris’ underlying data structures came thru.

Chris has built his technology on a couple of assumptions which are: IF is not storytelling with a touch of interactivity or interactivity with a tiny touch of storytelling. It should be a solid mix of both. But since interactivity is harder for us, we should make sure it is working first.

He sees a division of labor between the author of a story world, the computer and the user. The author creates the world, then gives control of the working out of the actual story (the ’instantiation’ of the story in programming terms) to the computer and the player.

I had not looked at the Erasmatron for about 3 or 4 years and Chris has made a lot of progress since then. He literally talked for 3 hours and just finished sketching out the features and ideas behind it so I won’t try to summarize everything he said. I will just list the highlights, and some reservations that I had.

* It is based on the laws of drama, not the laws of physics.

* The character’s in the game have unique personalities which respond to the story world events and their ’emotional memory’ of their relationships with you and other characters.

* They remember and respond to past events.

* Within the game there is a simple language that the non-Player Characters (NPC’s) use to talk to the player.

* The NPC’s will actually talk on their own, lie to each other, track down lies, gossip, and lie to the player of the game.

* In order to track down lies they will SPY on each other and the player!

Non-programmers may not realize this but this is absolutely amazing! People are burning the midnight oil building sparse IF with a hundred or two hundred nodes. But with Erazmatron HUGE story lines become possible with stupendous savings in time and energy.

* The biggest problem with the Erasmatron is that it does not have any front end to make it look like anything. Right now all of this stuff is hidden away from the user. Chris has been hoping that some one will build a nice graphical user interface for it, but with no commercial products out yet it has not happened. I suppose it is possible for someone to use this as a back end to their game. But I think this is unlikely. Right now, the game industry is hit based where the top game (in a handful of well established categories) gets tons of money and everyone else hopes to break even. In such a market people get very leery of innovating because being away from a clear category, almost assures that you WON’T get that top slot. The upshot is that Chris’ technology is in a Catch 22 situation. Without money to build a nice front end there are no saleable products and without nice products there is no money.

* My second reservation is more troublesome. This product is very different from anything else ever created. People don’t understand how to approach it. What is more, in movies or books there are moments of high drama. In Chris’ engine I suspect that the drama will be there but it will be understated and spread out. I suspect that you will have to pay attention to get the most out of the product.

* A final question (this may already be covered). How does Chris’ engine queue up dramatic music? I have looked at his API and I don’t see any interface for telling the front end that we need romantic music here or the climatic battle music at this time. If more than half of the dramatic impact from a game comes from its music, it seems a large oversight. (Again, I simply may have missed this.)

Laura Mixon-Gould built a storyworld in an earlier version of Chris’ engine, and it is basically half done. She briefly told us about her experience. She would like to take another crack at it and simplify her story-world to give it a tighter dramatic focus (and cut down on her labor.) She had a couple of suggestions for Chris that he has implemented and a couple of new ones. The features she would like are some built-in verbs that allow the player to track down chains of gossip. She would also like better support for the introductory sequence to let people know what was going on.

I came up with a suggestion for Chris’ technology myself: Someone had mentioned that the NPC’s (non-Player Characters) are so BUSY behind the scenes; talking and spying and gossiping and fighting each other; that it is hard to keep track of what is going on. My suggestion is that Chris missed a dramatic law: "Not much happens unless the hero or the villain is there." I think that there should be a good chance that a NPC would wait and do nothing when not on a stage with a mover and shaker.

We then did a round robin trying to find if there were things we could unanimously agree on. Chris summarized this on his brief report of the event. We ended up the session calling up Doug Sharp and en masse wishing him a ’get well soon’ / ’we missed you’. The group gave Chris a book in thanks for all his work.

All in all, a very successful trip for me.

Richard Smith