Racket

Why Racket doesn't have a uniform collection interface? It's fucking stupid to have string-ref, vector-ref, list-ref etc in the language.

scheme is for niggers

>>1
You can easily make your own.
http://docs.racket-lang.org/reference/sequences.html

But you'd use car/cdr with lists anyways.

>>3

(define (sequence-ref seq n)
  (let-values (((next? next) (sequence-generate seq)))
    (let loop ((r '()))
      (if (next?) (loop (next)) r))))
(sequence-ref '(1 2 3) 2) ; 3
(sequence-ref #(1 2 3) 2) ; 3
(sequence-ref "123" 2)    ; #\3
(sequence-ref #"123" 2)   ; 51 (#\3)

>>3
>>4
http://docs.racket-lang.org/reference/sequences.html#(mod-path._racket/stream)

According to the page you linked you don't even need to write your own. Or am I missing something?


(stream-ref '(1 2 3) 1)
2
(stream-ref #(1 2 3) 1)
2
(stream-ref "123" 1)
#\2

>>5
I thought they were delayed lists and jumped that part.
Yes, you can use that.

>>6
But the question still stands, why there are distinct functions for each type, why not to have just one generic ref, length, map that would work with any sequence?

>>7
You may want your code to be data-structure specific, and because of stream-ref returning two values when applied to an hash (they could just return a cons with (key . value)).
The non-polymorphic ``ref'' (that simply doesn't exist), length, map, etc are inherited from old Lisps, and freshmen in CL often wonder why they are not defined as generics.

However, you could just define your module/#lang that exports them as the stream-* variants and use it.

>>8
Also, forgot my name.

>>9
Look at this faggot!

1. It should be possible with swindle, unless they took that out of racket. (ditto for all the clos-alikes like coops & goops)
2. I seem to recall a collections SRFI, but it (obviously) didn't take off.

>>3
>>4
>>6
lol too sicp didnt fread