[plt-scheme] metaphors for local

On Nov 5, 2007, at 5:27 PM, Prabhakar Ragde wrote:

> Matthias Felleisen wrote:
>
>> Most of what you have mentioned is flawed.
>
> I'm sorry, I wasn't clear. I meant something with the metaphoric  
> power of the stack of books for a stack, or the cache of books for  
> a cache. I didn't mean that these metaphors applied to local.  
> Anyone coming from a primarily imperative background will, if they  
> are not careful, try the stack metaphor and screw it up. (Since  
> Pascal went away, the amount of experience with lexical scope out  
> there has diminished dramatically, even if gcc does support it,  
> sort of.)
>
>> I recommend to think of a pile of papers except that every  
>> programmer has his own secret ink with which to definitions that  
>> only he can read. Don't ever throw any of these papers away! You  
>> might need them.
>
> Yes, but these papers are written at compile time and read at run  
> time, and we don't have these notions when local is introduced in  
> HtDP. At least we don't have mutation, either. --PR

Pardon my mistake:

1. Even if you use secret ink at both times -- writing the program  
and evaluating it -- you don't get away with it. You need new secret  
ink every time you evaluate a local.

2. The return has nothing to do with set!. Just look at the d/dx  
example that I sent out for closures. Same problem.

-- Matthias