Smalltalk is a tidier functional language than Scheme

Mike Klein mike at twinsun.com
Wed Sep 2 04:38:58 UTC 1998


On Tue, 1 Sep 1998, Patrick Logan wrote:

>   In my mind, this is one of the niceties of Smalltalk over Scheme...
>   there are no special forms.  If the parameter is a block, it
>   evaluates to a BlockContext (or closure... hopefully, someday)
> 
> Well, Scheme doesn't require special forms. It's just that a closure
> in Scheme is less tidy than in Smalltalk!
> 
>   | z |
>   z := 0.
>   [ :x :y | z := x + y + z]
> 
> vs.
> 
>   (let ((z 0))
>     (lambda (x y) (set! z (+ x y z)) z))
> 
> Scheme has that wordy "lambda" keyword. Here's how a boolean based on
> closures would look in Scheme...

Of course, a reasonable implementation would actually use the greek lambda
symbol ;-)  (and the emoticons would have Unicode characters as well)

>   (define true (lambda (trueBlock falseBlock)
>                  (trueBlock)))
> 
>   (define false (lambda (trueBlock falseBlock)
>                   (falseBlock)))
> 
>   (set! boolean (compute-true-or-false ...))
>   (boolean (lambda () (do-this-if-true ...))
>            (lambda () (do-this-if-false ...)))

Yeah, I call these Church Booleans.

> Not as tidy as Smalltalk! Smalltalk is a nice *semi-functional*
> programming language! Objects and functions are *very* similar!

I agree, but look what a nice job they did on the numerics, not to
mention one of the best written language specs *ever*.

I wish we lived in a world where comparing Smalltalk and Scheme was a
major issue.  Unfortunately, on any spectrum that includes C++/java,
Scheme and Smalltalk are unresolvable close together.

Whistfully,
	Mike Klein





More information about the Squeak-dev mailing list