Question

These days I have took Peter Norvig’s Udacity course CS212: DESIGN OF COMPUTER PROGRAMS.
Unfortunately, the course is all in Python so, for the sake of learning, I wrote the equivalent code in Racket for regex compiler given in unit 3 of that course.

You can see my code here: http://codepad.org/8x0rMXOi

Now, what’s bothering me is that Mr. Norvig’s original code in Python is somewhat shorter than mine. 🙁 But ok, I’m just beginner in Racket, and that is expected. But I wonder if some Racket expert can shorten my code, so that Racket code become shorter than original Norvig’s Python code?

Answer 1

Here are a few tips.

The ormap-expression in:

 (define (in-chars? c chars)
   (ormap (lambda (ch) (equal? c ch)) (string->list chars))) 

can be written as

   (memv c (string->list chars)) 

The if-epression in

(define (match pattern text)
  (define remainders (pattern text))
  (if (not (set-empty? remainders))
      (substring text 0 (- (string-length text)
                           (string-length (argmin string-length 
                                                  (set->list remainders)))))
      #f))

can be written as

 (and (not (set-empty? remainders))
      (substring ...)

However your function are small and to the point, so I wouldn’t change much.

A more convenient syntax for manipulating strings would make
it easier to read and write string manipulation programs. Some years ago I made an attempt and wrote a concat macro.

I used it to implement Norvig’s spelling checker (his original article might interest you). The resulting spelling checker and the concat macro is explained here

http://blog.scheme.dk/2007/04/writing-spelling-corrector-in-plt.html

Update: I have written an updated version of the spell checker.
The concat macro is makes simple string manipulation shorter.

https://github.com/soegaard/this-and-that/blob/master/spell-checker.rkt