Tag Archives: concept

a little scheming

21 Sep

A fair while ago I had an excellent meeting with Alec Resnick; high time, then, to type up my notes from that. Since I wrote them on the subway back from Davis, they’re all out of order; this’ll be fun.

The little schemer is an old book/textbook for learning lisp, but what’s incredible about it (and useful as we consider how to motivate students to learn on their own) is the way it draws you into exploring lisp and forming a mental model just by reading it. The format is simple: the left column is a short, unambiguous question, while the right holds the answer and a brief explanation. The questions start with basic building blocks, like “is ‘a’ an atom?” [yes] and “what is the car of l, where l is (((hotdogs)) (and) (pickle) relish)”  [((hotdogs))], and the question-and-answer format make the reader form and continually test an internal model of the programming language. By the end of the first chapter, all of the axiomatic building blocks have been learned; the other chapters use them to build basic recursive function that then build other functions, and so on. What amazed me about the book is how much reading it feels like exploration, rather than learning; sure, it’s also very funny, but reading it is more like playing a game within scheme than it is like reading a textbook.

One of the things Alec said was that he’s seen many educational sites missing the feedback elements and student focus of PENed. As we start to write lessons for TTI students, we’re facing some difficulties on the feedback front; another note here raises the idea of making a TTI-specific wiki site that would automatically sync with an MIT-hosted one; if we can do this, then the idea of live in-lesson feedback from a webpage becomes much more feasible. Also, could we put a local copy of wikipedia, or at least the important articles, on there?

Alec raised an interesting question to think about when writing and planning material: what do the students not need to know, or have a good grasp of? What ends should be left loose?

Another question on a similar note: can we design lessons and experiments such that the fail in interesting ways? They could be designed to fail from the beginning (puzzle circuits, or they could be designed such that common mistakes are easily checked: a circuit, let’s say, where putting your 7805 regulator in backwards causes an LED to light.

Continuing the thread of puzzle circuits (which could be things, already built, that require a circuit to be placed on them to make them work), we talked about how bounding a design challenge (limiting the allowances that the student has) can make it less intimidating and friendlier to trial and error.

He also pointed out Star Simpson’s fuzzy circuits (plush circuit components that snapped together), which, like the many other circuit building blocks, are a way of reducing the mechanical requirements to building a circuit (especially with the fab lab’s primarily surface mount components) as well as bounding the problem, as mentioned above.

My last note here wonders whether the vark.com model (people ask questions, which get directed to relevant users, who answer them) could work well with the SF PEN club, and as a general way for us to let MIT friends help PEN: they could answer questions they know well.