I started with an intense 3-week introduction to Scheme (+1 week for SchemeUnit testing), working through the first nine chapters of HtDP. (Essentially, getting through structural recursion.) However, I didn’t take any time before then to really tie in 1. why we were learning Scheme, 2. why the recipe would pay dividends, and 3. how it would tie into writing interpreters.

So, in truth, the students were awesome that no-one bailed during those first four weeks complaining about the workload or wondering why it mattered. However, I could introduce that kind of “mastery learning” better, taking a day or two to talk more about languages in general, and providing the kind of context you’re describing.

It seemed that getting to objects (after first-class functions) was illuminating. In particular, the fact that objects are (essentially) just two environments (fields and methods) makes them look a lot like functions (closures). However, waiting until you’re in your fourth interpreter for that kind of “a ha” could, perhaps, be sped up or improved. Maybe.

Providing better context for the work was one of the suggestions. An idea here might be to compile to Javascript or Flash. Show the students that what they are learning can be used in the “real world” (and Javascript is an easy target language).