Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:61650
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: pass (pb1.pair.com: domain gmail.com designates 74.125.82.54 as permitted sender)
Mime-Version: 1.0 (Apple Message framework v1278)
Content-Type: text/plain; charset=utf-8
In-Reply-To: <CAAyV7nE6Of+SeDUGaJt2mNTT3RYJ4HdX_0PeL53x5doa6NWGow@mail.gmail.com>
Date: Mon, 23 Jul 2012 19:12:17 +0300
Cc: Andrew Faulds <ajf@ajf.me>,
 internals@lists.php.net
Content-Transfer-Encoding: quoted-printable
Message-ID: <9A437120-DD2E-4A75-A6D0-64B3E100DA83@gmail.com>
References: <2075060250.98513.1343049996761.JavaMail.open-xchange@oxwebmail.registrar-servers.com> <CAAyV7nE6Of+SeDUGaJt2mNTT3RYJ4HdX_0PeL53x5doa6NWGow@mail.gmail.com>
To: Anthony Ferrara <ircmaxell@gmail.com>
Subject: Re: [PHP-DEV] array_pick() - resent because email client broke formatting
From: g.b.yahav@gmail.com (Yahav Gindi Bar)


On 23 =D7=91=D7=99=D7=95=D7=9C 2012, at 19:05, Anthony Ferrara wrote:

> Andrew,
>=20
> On Mon, Jul 23, 2012 at 9:26 AM, Andrew Faulds <ajf@ajf.me> wrote:
>=20
>>=20
>> (resending because of broken formatting)
>>=20
>> Hi there,
>>=20
>> I apologise for my previous email. It was a disorganised mess that =
didn't
>> really
>> make the case for that function very well. So here's a proper =
proposal
>> (perhaps
>> too proper?).
>>=20
>> Introduction ------------
>>=20
>> I am proposing a function, named array_pick(), which takes a single =
(array)
>> argument, and returns a random value from it.
>>=20
>> Implementation --------------
>>=20
>> https://github.com/php/php-src/pull/142
>>=20
>> Specification -------------
>>=20
>> mixed array_pick(array $array);
>> The function returns the value of one random array element from =
$array.
>> Should
>> said array be empty, then it should return NULL.
>>=20
>> Rationale ---------
>>=20
>> - array_rand exists, however it only gets a random key, not a random
>> value. This
>> is the logical counterpart for API completeness
>> - array_pick is more convenient than $a[array_rand($a)], especially =
for
>> these
>> cases:
>>  - short array syntax + array_pick, e.g. =
some_func(array_pick(['foo.png',
>> 'bar.jpg', 'foobar.gif'])) vs $a =3D ['foo.png, 'bar.jpg', =
'foobar.gif'];
>> some_func($a[array_rand($a)]);
>>  - where using long names or nested arrays, e.g.
>> some_func(array_pick($foo['bar']['foobar']));
>> - array_pick is less wasteful than shuffling an array and taking the =
first
>> element
>> - $a[array_rand($a)] issuse a NOTICE from an undefined index if the =
array
>> is
>> empty, but array_pick() just returns null, which is more convenient
>> - I need this function myself quite often. It exists in two other
>> languages I
>> use that I can think of off the top of my head (Python, Game Maker
>> Language)
>>=20
>> Objections ----------
>> - PHP has too many functions
>>  - Yes, but that doesn't mean we can't add more. Otherwise we can't
>> improve
>> things and move forward.
>>=20
>=20
> Broken Window Theory. =
http://en.wikipedia.org/wiki/Broken_windows_theory
>=20
> Saying things are already broken is never a valid justification for
> breaking things further.
>=20
>=20
>> - This can be easily implemented in userland code/this is too =
frivolous to
>> warrant addition
>>  - This is true, however similarly to a lot of standard library
>> functionality,
>> it is used frequently enough that it is silly to have to rewrite it =
every
>> time
>> it is needed
>>=20
>=20
> It's a one-liner. It's not like it's non-trivial, or there are weird
> edge-cases, or security implications. It's a one line implementation:
>=20
> function array_pick(array $array) {
>    return empty($array) ? null : $array[array_rand($array)];
> }
>=20
>=20
>>  - It is also the logical counterpart to array_rand(). One for a key, =
one
>> for a
>> value.
>>=20
>=20
> Again, since it's so trivial to implement, why is it needed? Symmetry =
is a
> decent goal, but justified symmetry. Not just symmetry for the sake of =
it...
>=20
>=20
>> - The name isn't logical
>>  - Whilst you could argue that array_rand_key is a better name, I =
have
>> chosen
>> array_pick because it is shorter, and has a similar name to similar
>> functions in
>> Python (random.choice) and GML (choose)
>>=20
>=20
> Actually, I think that array_pick() is a poor name. Python's name =
makes
> sense, because it's bound to the random namespace. But being bound to =
the
> array namespace indicates a different thing. I would argue that
> array_random_value() would be a better choice (or any one of a number =
of
> names)...
>=20
>=20
>> ---
>> Thoughts?
>>=20
>=20
> Shared.
>=20
> Thanks,
>=20
> Anthony

> It's a one-liner. It's not like it's non-trivial, or there are weird
> edge-cases, or security implications. It's a one line implementation:
>=20
> function array_pick(array $array) {
>    return empty($array) ? null : $array[array_rand($array)];
> }

array_rand() is also trivial and one length long function, but it exists =
in the language and it is really useful.
I don't think that one should think about "how much line it takes to =
write it", you should think about "does many people use this function =
quite often in order to make it part of the language".=20
Personally, I think that this function is great, and when I read for the =
first time about array_rand() I thought it'll returns a value, after I =
realized it returns a key, I looked for a value pair and surprised to =
see that there's no value pair.