20 minutes ago, Rodolfo Carvalho wrote:
> It is possible to replace a pattern like this:
>
> (cond
> [..a..]
> [else (cond
> [..b..]
> ...)])
>
> With this simpler:
>
> (cond
> [..a..]
> [..b..]
> ...)
Speaking about such transformations and about things like (if E #t #f),
I have some code which I run against student submissions to find style
problems like these, and using syntax transformations means that it's
easy to make it actually suggest the transformation. For example,
feeding it this code:
(define (foo x y)
(cond [(< x y)
(printf "~s is smaller than ~s\n" x
y)]
[else (cond [(> x y) (printf "~s is bigger than ~s" x
y)]
[else (printf "They are ~a\n"
(if (if (eq? x y) #f #t)
"identical"
"equal"))])]))
produces this annotated printout:
(define (foo x y)
(cond [(< x y)
(printf "~s is smaller than ~s\n" x
;;> A
;;> (A) this expression (and the rest) should be on a separate line (or
;;> make the whole `printf' fit on one line)
y)]
[else (cond [(> x y) (printf "~s is bigger than ~s" x
;;> A B
;;> (A) flatten this into the surrounding `cond'
;;> (B) make this form fit on one line, or put it on a separate line
y)]
[else (printf "They are ~a\n"
(if (if (eq? x y) #f #t)
;;> A
;;> (A) `if' not needed, just use (not (eq? x y))
"identical"
"equal"))])]))
;;> A
;;> (A) misleading "flat" indentation
Actually, there's a few more things it prints -- it's not intended to
be used without a human going over its output. In case someone is
interested in such a project I can send out the code. It could even
make a cute drracket tool that criticizes your code...
--
((lambda (x) (x x)) (lambda (x) (x x))) Eli Barzilay:
http://barzilay.org/ Maze is Life!
____________________
Racket Users list:
http://lists.racket-lang.org/users
