Sylvia links to the Moving Learning Games Forward white paper from MIT, whose intended audience seems to be funders, particularly foundations ("The Role of Foundations" is the climax of the paper). I have, as you might imagine, a number of nits I could pick about this paper, which I'll keep to myself in the interest of your and my time. They do give "open"-ness more attention than most of the works in this genre, although they do it with all the verve and clarity of a student completing a punitive essay in detention hall, e.g. (p.44-45):
Open Source Engines
This idea takes the engines built on open standards and APIs a step further by open sourcing the code to the game engine itself. This provides a low cost way to both develop and distribute educational games, avoiding (for the developer) the purchasing of engines that can be very costly, and further avoiding (for the end user) that cost being passed along to them. Open source products have been a remarkable success in particular success – the most famous of which is the Linux operating system, but also includes applications like the Firefox web browser. Several open source game engines already exist (e.g. Ogre), but their popularity has been limited, perhaps by the tradeoff between price and functionality/ease of use, which has suffered in many open source products.
Pros and cons – One may ask the question about why such a product (or products – game engines for different genres) don’t already exist. Perhaps (like Firefox) they need a formal organization/foundation that can not only sponsor their initial development but sustain the community of developers and users. It may also be, from the commercial games industry perspective, that such an approach is unnecessary – that most of their development goes into assets and novel additions to game play that this approach would not help. But such an argument is not likely to come from the learning games community, where many developers would likely be happy to build on a solid set of shared features, concentrating on asset development, and content, contributing code when they are required to change the engine itself. However, unlike Linux, the end user community (teachers and learners) are unlikely to care about the product being based on an open source engine vs. something that is merely free (and closed source).
The very tricky thing they do here is restrict the discussion of "open source" to the "engines." That is, they completely elide the more important question of how the game itself (that is, the work done by the game development team, as opposed to its technical infrastructure; or put another way, the work the funder is paying the team to do) is licensed. It is great if both are free software, but of the two, I'd rather have the game logic be open source than the engine.
I don't have time to go on and on and on about this, but here's my recommendation to foundations and the government:
Require any software written under your grants be released as free and open source software.
The grantees will figure out the rest. They're not going to quit and go home or take up another occupation. They're smart guys, they'll make it work on those terms, and you'll get much better value for your investment when their work can be freely used, studied, modified and redistributed.
It is that simple. High expectations! No excuses! Quit coddling MIT developers!