Slashdot links to the first report of one of the MacArthur Digital Media and Learning grants ending with a whimper. The central problem in this case seems to have been simply underestimating the difficulty of creating a contemporary mmorpg. It is a little like a grant proposal on studying gender roles in cinema that starts with shooting a feature film. It isn't very practical.
They are, after the fact, making their work publicly available.. Of course, it is dependent upon a proprietary engine and proprietary tool chain, and they simply say nothing about the license, so it is all pretty unappealing and amateur-hour.
I'm not sure how this could have been avoided, other than being realistic about the economics of gaming. The visionary strategy would have been for MacArthur to use their money to bootstrap the open source infrastructure that is needed to really go forward with games in education, basically requiring that all the projects they fund be built on 100% open source, and encouraging the projects to collaborate on the same environments -- so that instead of each grantee writing their own game, they'd write one codebase with mods for different research questions. Whether or not that would actually work, I don't know.