Re: vector-find-median! and side effects William D Clinger 28 Aug 2018 02:41 UTC

Shiro Kawai wrote:

> If "fully sorted" is indeed the original meaning, I thing we shouldn't
> change it since it changes the spec.

I'm pretty sure "fully sorted" was the original intention, because
the sample implementation of vector-find-median! is preceded by this
comment:

;;; For this procedure, the SRFI 132 specification
;;; demands the vector be sorted (by side effect).

I think it might be useful to add a post-finalization note emphasizing
that vector-find-median is likely to run quite a bit faster than the
imperative vector-find-median! alternative.

Will