At the postmortem earlier this week, one of the presenters was Chris Canfield. I presume this is him, although the page contains precisely nothing in common with his presentation.
One of the things he's doing on the side is building chatbot-based games. The basic idea is that higher fidelity - graphics, better graphics, insanely good graphics - are not only more and more expensive, but are also more and more obviously fake. So scale back down, you get a more believable (and cheaper) experience.
Also, he used the phrase "burn content". It makes me feel all happy to hear it tossed off like it's an obvious, well-understood concept.
Anyway, the basic idea is not terribly new. Text games are cheaper to make, if you like gross understatements, and many people are pleased by the kinds of content you can have. For example, it's very difficult to include thoughts or indirect information in a visual game. How do you say, "he knew she was carrying the photo with her even after all these years" visually?
Text adventure games are, however, a limited genre. Consisting of X parts inventory management puzzles and Y parts reading the designer's mind, calling them "interactive fiction" is an arrogant stretch. There are a few games out there which attempt to be something else, usually something involving characters that react realistically, but they are generally pretty haphazard, opaque, and confusing.
Chris' approach has been to separate the games out into what is functionally two games. One game is an information game: going to websites, reading fake blogs, collecting information that you can assemble into a picture of what's going on. The other game is a natural language chat which serves as the hub for the information game and is how the game is actually "won".
For example, in a murder mystery, you would chat with "guests" to get some information, follow the leads on the internet, chat with guests about what you discovered on the net, and so forth until you found the guilty party.
In some ways, this is a great approach. It allows a game world to exist without any actual need to have it implemented in an engine. You not only don't need a 3D model of the city, you don't even need a text adventure node map. The only way you encounter the world is through the "eyes" of one of its inhabitants, talking to you or posting about it on the internet.
This has a really, really powerful upside on the other half of the equation, as well. Because you are seeing everything through the eyes of the characters, you will really get a feel for the characters. Presuming the chat phase isn't so hideous that it breaks immersion, this is a great way to build a deep interest for the characters in a game.
From his description, his examples all focus on a heavily scripted setup. The games play through in one particular way - or you lose. They're oldschool rails puzzle games. This makes perfect sense because it means (A) you can build content instead of generating it algorithmically and (B) your chat bots don't have to adapt.
To me, that kind of play is pointless. Even if you script in multiple paths, it's not really very interactive. I like adventure games for what they are, but I need interactive systems to really get into a game. That's why I liked the Quest for Glory series so much more than the King's Quest series.
But once you create an interactive system, your content needs to be linked to it. If you can pick stuff up, then you need to have a system for inventory and what content can be picked up and what happens when something is brought somewhere and used.
Which basically throws the advantage of "chat-bot adventures" right out the window, since the whole point was to replace the complex, bulky system with a set of viewpoints.
Anyway, I have some ideas on that, but I've got to think them through a bit more. Just thought I would share the basic idea.
What's your take on text adventures?