Is it possible to have to have reducers in newlisp too as mentioned in this //http://adambard.com/blog/Reducers-explained-through-Python/ ?
The key takeaway from the article is
Quote
This is the idea behind reducers. If you take your mapping function (map, filter, flatten, etc.), and have it modify the reducing function, you can perform any number and combination of mappings without having to
Thanks, that is a very cool URL.
I sure could have used Reducers when I implemented the (combinations) function.
To make it efficient, I had to switch from map to dolist to avoid that exact problem that the Reducer functions address.
http://www.newlisp.org/index.cgi?page=Code_Snippets
(define (combinations n k (r '()))
(if (= (length r) n)
(list (sort r <))
(apply append (map
(fn (x) (combinations n ((+ 1 $idx) k) (cons x r))) k))))
; (combinations 2 '(a b c))
; => ((a b) (a c) (b c))
I'd really like to know how that would look using Reducers instead of dolist. dolist did do the job very efficiently, map blew up my computer.