From: gdefty@...
Date: 2008-08-07T18:35:28+09:00
Subject: [ruby-core:18164] Re: New array methods cycle, choice, shuffle (plus bug in cycle)

In message "Re: [ruby-core:18133] Re: New array
methods cycle, choice, shuffle (plus bug in cycle)"

On Thu, August 07, 2008, 14:48:00 Yukihiro
Matsumoto wrote:

|    on Tue, 5 Aug 2008 23:37:57 +0900, "Martin
DeMello" <martindemello@gmail.com> writes:
|
||The second point looks wrong to me - I'd expect
sample! to remove a
||random set of elements from the array without
disturbing the ordering
||of the remaining elements.

|Then what your expectation works for?  I cannot
think of a situation
|where the difference means much.  Is there any
use-case for it?

I have the same expectation of the order being
preserved. It is not because I have any particular
reason to preserve it, but simply because that is
how I would expect a reasonable implementation to
work.

The problem with the shuffle-and-shift solutions
is that they do not scale well with the size of
the array. It is fine for small amounts of data
(e.g. the rock-paper-scissors example) but not how
I would expect a built-in method to be coded,
where the size of the array can not be known in
advance.

If I take a sample of one or two elements out of
an array of thousands of items, I would expect to
generate one or two random numbers, not thousands.
(And the next time a sample is taken, generate
thousands more, because more items may have been
added to the array since the last sample).

It is in such a case that the difference becomes
evident.

Regards,

graeme