https://people.eecs.berkeley.edu/~bh/hiding.pdf https://people.eecs.berkeley.edu/~bh/hiding.pdf 7. On the other hand, we are about to unveil a radical unhiding

https://people.eecs.berkeley.edu/~bh/hiding.pdf https://people.eecs.berkeley.edu/~bh/hiding.pdf 7. On the other hand, we are about to unveil a radical unhiding: showing users the innards of expressions and procedure bodies via macros and metaprogramming. The central feature making this possible is the ability to convert back and forth between executable code (the blocks and scripts that we've always had) and syntax trees: lists of lists of the individual blocks and constant values (such as text and numbers), the building blocks of expressions and scripts. That conversion overcomes the only weakness of block languages, namely, that programs aren't data, which makes them harder to manipulate programmatically. 7½. Pedagogically, this is an extension of the self-reflection by means of non-hidden continuations, which call attention to the sequence of events in executing a script, and which we inherited from Scheme. Continuations are conceptually simple for the implementor, since they already exist in any interpreter for any language, and it's just a matter of making them visible to users. But they're not conceptually simple for users! The fact that they can be called repeatedly, and from outside of the script in which they were created, feels like magic. And it is; it's the magic of "everything first class." But the point here is that we didn't add this feature in response to a specific pedagogic need. Rather, explicit continuations help advanced programmers write advanced programs, and a few library blocks. And they're a way to plant the flag of Scheme in Snap!, which was a big part of our motivation.