On Nov 29, 2010, at 6:56 PM, Gregor Kiczales wrote: > In merging the mutable variables part of 1e with the universe of 2e, one > thing we do is move world state into a state variable. Doing that is nice > because it shows a classic use of mutable variables.
Exactly. And it is even better to do so in an OO context, where inheritance is used to override the default methods of the big-bang class (HTDC does this in some sense). > But its awkward to have to tell students something like "to keep big-bang > happy we have to let it pass around something, lets let it think the world > state is 1". The key is for the instructor to provide a teachpack that is appropriate. > I'm wondering whether anyone tried making a big-bang! in which big-bang > doesn't pass (or expect to receive) world state? I'm not asking about > implementing it since that's easy. I'm asking about whether anyone has tried > doing it and switching to that when they switch to state variables? I did so a year ago for the MS-level HtDP course. Either way -- set! based or class based -- works like a charm. This year's instructors are doing it too. Holler if you need help equipping the teachpack with proper error checking etc. -- Matthias _________________________________________________ For list-related administrative tasks: http://lists.racket-lang.org/listinfo/users
