Well, a really interesting evaluation would be to develop a pre-test/post-test for some learning objectives about what it means to understand code, and then hav

Well, a really interesting evaluation would be to develop a pre-test/post-test for some learning objectives about what it means to understand code, and then have this be the presentation of an instructional unit. You could measure this fairly empirically and simply. Of course, you'll end up having the question - "does this teach better than something else?"