Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:84554
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: pass (pb1.pair.com: domain gmail.com designates 209.85.212.173 as permitted sender)
Message-ID: <55006B7E.1080500@gmail.com>
Date: Wed, 11 Mar 2015 16:21:18 +0000
User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:31.0) Gecko/20100101 Thunderbird/31.5.0
MIME-Version: 1.0
To: internals@lists.php.net
References: <CA+FrEoKBhRErj2i_rSyeOyTBHAYw-OxjMKPGz6dDhayCMeWsKQ@mail.gmail.com>	<CAG1chorE4VwMJ0d=hs5pf8T-5z915fwv9uJvkRQ-NNnhKJPzaA@mail.gmail.com>	<CAMc0WS6SVsnQQ6pSzUeCpDRzEwBZFeos1CwXMxxzHF3YjR5txA@mail.gmail.com>	<etPan.54fda4b0.333ab105.941f@Mikes-iMac.local>	<CA+FrEoL1tOuExW8wPGiO1kMa8pGa72NR3zgxsOwUuS0bcr4BuA@mail.gmail.com>	<etPan.54fda677.75a2a8d4.941f@Mikes-iMac.local>	<CA+FrEoJJyKh7x_N0ojL7mgCNgN+3RnS7pCS58rXYoWiUz8MNXg@mail.gmail.com>	<op.xu8tk4pkhuszwh@tsukapiou> <24.71.03288.592FDF45@pb1.pair.com>	<CALW=WEAud32Pbvq8xMZT0YDRBwHrEvk4=NhKdv6jMPB2gj=-fw@mail.gmail.com> <CF.50.23562.56A0EF45@pb1.pair.com> <54FE20E5.1060605@gmail.com> <A3.0C.08808.8385FF45@pb1.pair.com> <55003D12.5010006@gmail.com> <AC.B3.07702.AF440055@pb1.pair.com>
In-Reply-To: <AC.B3.07702.AF440055@pb1.pair.com>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [PHP-DEV] Request Feedback for Instance Variable Sugar RFC
From: rowan.collins@gmail.com (Rowan Collins)

Johannes Ott wrote on 11/03/2015 13:36:
> Am 11.03.2015 um 14:03 schrieb Rowan Collins:
>> Johannes Ott wrote on 10/03/2015 20:46:
>>> okay indeed the dynamic properties are a problem I didn't think about on
>>> my suggestion. Without the wish to hijack this thread for another
>>> typesafety discussion, I must say again that PHP needs a less dynamic
>>> and more declaratively properties concept in my opinion.
>> Yes, I think a standard way to say that a particular object has strictly
>> declared all its properties, and should not allow  would be useful.
>>
>> (Even with that, though, I'd be against guessing that "$foo" meant
>> "$this->foo"; it just makes code harder to read.)
>>
> I disagree, programming Java beside PHP since about 15 years now,
> personally I think always having "this"-keyword, where it is not
> necassary in a strict OOP-world, makes the code more unreadable for the
> simple fact it is more to read.

My reasoning is that code that is ambiguous is hard to read. If "$foo" 
can mean either "a local variable called $foo" or "a property of the 
current object called $foo", then you have to know which it is in order 
to understand what code is doing.

This is not about "a strict OOP-world", incidentally, it's about scoping 
rules. Java imports object properties into the scope of each method, PHP 
does not. Even if properties had to be declared (which is probably a 
good idea), local variables still wouldn't be - in Java, the fact that 
it's not declared locally means it *must* be coming from somewhere else.

I also know that when I was first learning Java at school, it confused 
me immensely which variables I was allowed to access in static contexts. 
In PHP, that's simple - if you can't access $this, you can't access any 
of it's properties.


>>> So I would suggest for now to keep the $this variable, but to make it
>>> more similar to other OOP-languages I would suggest to remove the
>>> $-character in front. In my opinion it would fit better to other object
>>> keywords like parent and self as well.
>> Other OOP languages only don't have a sigil such as $ in front of this
>> if they don't have one in front of *any* variable. Why should $this,
>> which acts like a normal variable in pretty much every way, drop the $
>> when every other variable in the language has one?
>>
>> Note that the syntax of parent and self is different, and is consistent
>> with static member/method access, or more strictly "scope resolution".
>> You can't pass parent or self around as variables, only use them to
>> resolve scopes, so they don't have a $.
>>
> I only agree a bit, because the keyword "this", is what to call a
> hybridization, more often used to define the scope of the property or
> method you want to use, then used really as a pure variable.
>
> For example:
>
> this->a;
> or
> this->some_function(...);
>
> just defining the scope as in the current instance is much more used then.
>
> some_function($this);

This is true of any object variable.

These resolve the scope of the property/method to the variable $foo:

$a = $foo->a;
$foo->some_function(...);

These resolve the scope of the property/method to the variable $this:

$a = $this->a;
$this->some_function(...);

If you want to resolve the scope of a method to the current object 
without using the variable $this, you can also use the "static" keyword; 
these are equivalent:

$this->some_function(...);
static::some_function(...);
// http://3v4l.org/E0XYs

There is nothing unvariable-like about $this, so if variables begin with 
$, $this should begin with $.

Regards,
-- 
Rowan Collins
[IMSoP]