Quote:Original post by Alpha_ProgDes
*** Source Snippet Removed ***
i take it "we" is basically a new object every time i redefine it. something like:
*** Source Snippet Removed ***

i'm using Dr. Scheme by the way.

I'm not 100% sure what you really mean. There are multiple ways to answer that question.


'define' (re)binds the identifier to a value. If you write this in Scheme:

(define x 5)(define (f) x)(f) -> 5(define x 6)(f)--> 6

However, you cannot use this to achieve runtime mutability: if you use 'define' in a new environment/scope, the old value will be restored:

(define x 5)(define (f)   (define x 6)   'done)x--> 5(f)--> donex--> 5

The 'define' in 'f' creates a new environment (scope limited to f's body), in which a new 'x' is created.

To show this is not a simple matter, consider the following O'Caml code:

let x = 5;;let f () = x;;f ();;--> 5let x = 6;;f ();;--> 5

Don't break your head over it for now, chapter 4 will make this more clear where you'll write your very own Scheme interpreter. I would suggest you don't think of 'define' as a way to redefine values; act as if it's only allowed to use one define per identifier per scope.
edit: I see that Scheme (DrScheme anyway) prevents you from 'defining' the same identifier twice in an environment which is not the top-level one, which is good.



There's also the "referential transparency" issue. In Scheme, everything works through pointers/references. If you were to work in C++ this way, you could have two different pointers to 5, without them being equal because they point to different memory cells:

int a = 5, b = 5;int* p = &aint* q = &bp != q

As far as I know, there is no way to make this kind of distinction in Scheme with integers. Some languages take this approach everywhere: if you have two data structures containing the same data, they are indiscernible from each other. Only "object value" counts in these languages. Other languages such as C/C++/java/C#/... take into account "object identity": objects can represent the same values, yet have different identities (hence you need to use .equals() instead of == in java, etc.)

Quote:Original post by Alpha_ProgDes
I have a semantics question.
*** Source Snippet Removed ***

COND may have an arbitrary number of consequent expressions/commands for any particular predicate. For example, (cond ((> x y) (draw x) x) (else (draw y) y)) is valid. To do the same with an IF statement one would need to wrap the sequence with BEGIN. I'm not sure what the motivation was for the COND syntax... I guess they felt it was clearer to wrap a whole predicate/consequent clause in parentheses than have everything lifted up a level and have the meaning of terms in COND expression depend on some index.

Quote:Original post by Alpha_ProgDes
Okay Section 1.2 problem 1.10 is killing me!! I figured out the first two but I can't figure out (A 2 4). It goes 2 4 16, then 65536! What kinda wacked out function does that!!!!?????

Since this hasn't much to do with functional programming, feel free to skip the exercise.

If you're interested in the solution: you know that

• A(1,0) = 0


• A(1,n) = 2n if n > 0.

If you use this in the expansion of A(2,n) you get

A(2, n) = A(1, A(2, n-1))        = 2^A(2, n-1)        = 2^A(1, A(2, n-2))        = 2^(2^A(2, n-2))        = 2^(2^(2^A(2, n-3)))        = ...

So, using Knuth's up-arrow notation, we get A(2, n) = 2 ^^ n for n > 0:

A(2, 1) = 2 ^^ 1 = 2
A(2, 2) = 2 ^^ 2 = 22 = 4
A(2, 3) = 2 ^^ 3 = 222 = 16
A(2, 4) = 2 ^^ 4 = 2222 = 65536

Quote:Original post by Alpha_ProgDes
Alright I'm reading the notes and I came across this.
*** Source Snippet Removed ***
When will I have or even want to use the lambda expression over the shorter define expression? Or is it just syntactic sugar that I shouldn't worry about?

The first is sugar for the second. You would usually use the first form, but if you wanted to curry a function, for example, you would have to use the second form.

Quote:Original post by Alpha_ProgDes

Quote:Original post by Rebooted

Quote:Original post by Alpha_ProgDes
Alright I'm reading the notes and I came across this.
*** Source Snippet Removed ***
When will I have or even want to use the lambda expression over the shorter define expression? Or is it just syntactic sugar that I shouldn't worry about?

The first is sugar for the second. You would usually use the first form, but if you wanted to curry a function, for example, you would have to use the second form.

Curry?

Currying comes up in one of Chapter 2's exercises. We covered it a little here.