On Jan 4, 2015, at 7:52 AM, Gustavo Massaccesi wrote: > The original code is something like: > > (define ALPHABET (in-list (string->list "abc...xyz"))) > (for/list ([c ALPHABET])) > (string-append "?" (string c) "?")) > > The first problem is that outside a clause of a `for` the `in-list` > creates a sequence. But inside a clause of a `for` the `in-list` is > spliced and generates more efficient code. > > (define ALPHABET (string->list "abc...xyz")) > (for/list ([c (in-list ALPHABET)])) > (string-append "?" (string c) "?")) > > The second problem is that for each word `(string c)` creates a new > temporary string, so I mapped `string` at the definition site. > > (define ALPHABET (map string (string->list "abc...xyz"))) > (for/list ([c (in-list ALPHABET)])) > (string-append "?" c "?")) > > --- > > Following the idea of the second change I tried to remove more of the > temporary strings in the code ( for example, `(substring x 1)` looks > like a bad idea). But my tries were unsuccessful to make the code > faster :(.
I had experimented with these on my Mac and decided that it didn't matter. A near-10% improvement is not what I saw so thanks for the correction. -- Matthias ____________________ Racket Users list: http://lists.racket-lang.org/users
