At UT Austin, we've been building a 3-D interactive environment called Nero [1] for some time now. The project's goal is to have an all-in-one, visually appeali

At UT Austin, we've been building a 3-D interactive environment called Nero [1] for some time now. The project's goal is to have an all-in-one, visually appealing platform to teach and explain different concepts in AI. It was started as a closed-source collaboration with a games company and has since been rewritten as an open source project. I've been contributing to the codebase for the last few months, so I have some experience here. First off, teaching anything non-trivial is difficult. By non-trivial, I mean formal knowledge that requires you to actually have a deep understanding of the material, as opposed to rote learning or reusable strategies like most games involve. I haven't played Skyrim, so if there is some actual expertise you acquire through playing it, I'd be happy to hear about it. Imagine you're trying to teach statistics. You start out with some gambling game to teach basic odds and probabilities, and that's great. Now how does that transition to Gaussian distributions and standard deviations, one of the most fundamental and comparatively simple concepts in statistics, without simply being boring? It's not so clear. Second, is there any real money in it? I suppose if you can get lots of schools to use your game, then great. But of course we know how long those sales cycles are, and you can't really expect your game to look all that awesome in 3 years after you've finally worked your way up the organizational ladder and closed. So you probably need to either sell your game directly to consumers or give it away. If you're going to sell it to consumers, you have to compete with other games in your market. This is tough because a game that requires little critical thinking will likely be more appealing to a broad audience. Again, maybe Skyrim requires truly critical thinking along the lines of understanding derivatives and integrals, and if it does please correct me, but I doubt it. Most people buy videogames to play in their downtime, so they want to relax. This means your market will be very niche and I just do not see the long-term potential. The other option, giving it away, puts you in a similar situation to what we're doing at UT. If you are giving it away, you probably need to find people who will work for cheap because grants do not end up paying much for this sort of thing, enter students. Our situation is such that we have had literally dozens of people adding, changing, and branching the code with varying degrees of quality in commits (looking at you, undergrads), creating lots of technical debt. We also are not experts in game development, so the graphics and mechanics are a little rough. That said, it's a pretty decent platform for teaching and demoing different concepts in AI. But is it as awesome, polished, and fun as Skyrim looks? Of course not-- you get what you pay for. I guess my main point is that there are plenty of legitimate reasons that such teaching environments are not pervasive in modern education. It's one of those ideas that is easy on the surface but quickly breaks down in practice. When it works, it makes for neat demos, but to make it work at the same scale as a commercially successful video game just seems infeasible in today's market. [1] http://opennero.googlecode.com http://opennero.googlecode.com