The projects are pitches from real local industry/government/research organizations. There was on instructor responsible for the class that meet with the groups from time to time, milestones basically, but by and large, the students were on their own. And not only was it useful because we had to apply everything we’d learned the three years previous, but you had to seriously, earnestly do real team-based work. You had to understand the dynamics of working on a team, handling other personalities, etc.

It would have been fantastic if small projects like that had been part of the whole curriculum; given how useful the senior project was, smaller versions of the same would have probably enriched by undergrad education quite a bit.

I lot of what you suggest rings true to solid educational theory. You are hitting all around the notion of Vygotsky’s Zone of Proximal Development which basically states that people only learn things when they have enough exposure to the necessary pieces of a complex task to parse, order and accomplish that task. An educator’s job is to ‘scaffold’ the student so that they can do more than they are able to do on their own, and to slowly take away the scaffolding as their understanding increases so that they can complete the task completely on their own.

Of course by that time the next task is also underway and the scaffolding is in place for this new, more complex task.

In this regard, structuring a curriculum is an exercise in game design in a high stakes game. And the iteration time is really quite slow for making changes/improvements. Also, the organizational structure of schools actively works against failing fast/rapid iteration. However it also serves to prevent a ‘chasing the latest tool’ fetish that could emerge.

I will be reading and re-reading this over the weekend. A lot to digest. If you wouldn’t mind, I’ll follow up with you in 140 character chunks or via email.

This post was mentioned on Twitter by fuzzybinary: New Blog Post: Marketable Skills: State of Game Programming Education. http://bit.ly/xmBCH Comments welcome….

Theory classes tend to contain way too much code when it comes to learning data structures or basic concepts related to code. Classes that teach concepts like grammars and discrete math are not ever applied throughout the curriculum and most students left those classes wondering what their use was. Performing a true split between theory and practice — but balancing them so that theory is always applied would help retention and perhaps even appreciation of the theory.

As for the inner workings of modern chipsets, cache coherency, synchronization primitives, compiler intrinsics, and virtual memory? You are quite right, in my CS program they only taught us synchronization primitives and virtual memory. Very little time indeed was devoted to the compiler or actual hardware we were operating on. At least my university had yet to transition to Java by the time I had graduated, or else I would have never gotten much C++ experience in. (although it was my C# and XNA knowledge that landed me my job)

The real reason I didn’t go to a game school was because I was afraid that with how the recruiting booths are at GDC (they treated new grads like lepers from what I saw) I didn’t want to count on getting a game job right out of the barrel and thus wanted a more “generic” education with my own private game studies on the side. All in all it worked out well for me, but I imagine many other students don’t have the luxury of devoting the time they should spend doing homework on learning various graphics APIs and playing with game toolkits and engines…

I am afraid I am starting to rant and should stop here.