This post is almost a direct response to Andrew’s post on GTxA, in response to Janet’s response on Andrew’s panel at GDC. As Andrew point’s out, the responses from the panelists are available.

Where I disagree with Andrew is in his comments on interface in the first point. I’m not convinced that natural language is the best idea as an interface for game based interactive stories. In IF or purely interactive drama, I might agree, but there are many problems with such an interface. Interestingly, GTxAuto posted parts of a response that I agree with a few days later. What’s interesting is that Andrew doesn’t quote what I feel is the most important part of the original post, where the real problem with natural language systems is revealed.

Frustration happens when interactions fail. Either the participant didn’t know enough to form expectations as to a result (illiteracy), or the interface led the participant to form bad expectations (inconsistency). Either way, the participant guessed, and guessed wrong. Guessing leads to failure, and failure leads to frustration.

For those familiar with my work, I refer to this as the “Willing Suspension of Freedom” (a term I believe I made up since I can find no mention of it anywhere ot the web), something I believe is central to all interactive works and applications, not just games. Both that and the willing suspension of disbelief need to work together in order for an interactive story to work as a whole. In my opinion, natural language is always going to cause involuntary breaks in the willing suspension of freedom, no matter how good the parser is.

What it comes down to is this: using natural language processing makes people believe that anything typed should get a proper response, and although this can be true at least 50% of the time, from a game perspective this is not good enough, especially when it’s understood that games are based on rules and regulations. In the case of most mass market games, adding “invisible walls” isn’t necessarily a problem (despite rants to the contrary) so long as it’s done well and thoughtfully as part of the design, instead of completely arbitrarily or because of technical limitations.. The loss of complete control is okay, so long as it’s consistent with my understanding of how the game should work.

Instead of adding more verbs as both Andrew and Chris argue for, I am now firmly of the opinion that our verbs simply need to be more meaningful. For example, walk and look are very simple verbs that are in almost every modern computer game, but where you walk and where you look are not usually considered meaningful. Why not? In real life, if I walk up to a bar and look directly at the bartender, eventually, he’s going to come down and ask me if I want a drink. If I walk up to a bar and look directly at a woman sitting there, she’ll probably look back and (depending on her character) start a conversation (“May I help you?”). Here, we’ve added no verbs, just made verbs already freely available more meaningful. From a game perspective this doesn’t complicate the issue for the main stream audience. I’m still all for adding verbs, so long as their addition doesn’t overwhelm the interactive story and the game you’re trying to make.

This probably came out all wrong, but hey, at least it’s a response of some sort. Also note that I’m pretty sure I can’t go a single blog post without linking to my thesis.