Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:14828
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Message-ID: <5.1.0.14.2.20050211084917.02183590@localhost>
Date: Fri, 11 Feb 2005 08:50:09 -0800
To: Ante Drnasin <ante.dfg@moj.net>,internals@lists.php.net
In-Reply-To: <20050211084054.4290.qmail@lists.php.net>
References: <1108070462.20077.129.camel@localhost.localdomain>
	<20050207141916.21665.qmail@lists.php.net>
	<1062509740.20050209103014@marcus-boerger.de>
	<20050209110525.1936.qmail@lists.php.net>
	<20050209164713.GC25360@gravitonic.com>
	<20050210080546.52065.qmail@lists.php.net>
	<20050210185558.GA64423@gravitonic.com>
	<1108068700.20077.121.camel@localhost.localdomain>
	<1108070462.20077.129.camel@localhost.localdomain>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
Subject: Re: [PHP-DEV] Autoboxing in php 5.1
From: andi@zend.com (Andi Gutmans)

I think phpDoc is the solution because especially as we start going into 
the web services realm we're going to need to document more complex 
signatures and this kind of syntax won't be sufficient anyway.
We will need to have a way to document web services so that we can 
auto-generate WSDL files.

Andi

At 09:43 AM 2/11/2005 +0100, Ante Drnasin wrote:
>John Coggeshall wrote:
>>To be clear:
>>Although I think this might be implemented as some sort of object I am
>>not interested in making objects out of everything. All I want is this:
>>function foo(Integer $a, Float $b, String $c, Boolean $d) {
>>}
>>and be able to introspect against that... how that ultimately gets
>>implemented (likely PHP would auto-cast that to an int) doesn't matter
>>to me... but the point here is I think the ability to have more
>>*optional* meta-data associated with function declarations shouldn't be
>>brushed off as nonsense when there are real benefits to the ability.
>>John
>>On Thu, 2005-02-10 at 15:51, John Coggeshall wrote:
>>
>>>>You did't answer my question. Why?
>>>
>>>I am def. a fan of this idea. I'd love to see internally a set of
>>>Java-style objects representing the basic types in PHP.
>>>As for why I have two reasons:
>>>
>>>Although PHP is not a strongly-typed language and never will be, with
>>>the introduction of type-hinting I feel that having a standard set of
>>>objects representing the fundamental types in PHP will allow developers
>>>of libraries to enforce the proper restrictions if they desire on the
>>>types which end up in their functions. This further degree of control
>>>over the architecture in a PHP class makes an architecture tighter and
>>>easier to manage with growth.
>>>
>>>More importantly than the concept of type-hinting, the lack of typing
>>>information is holding PHP back signifcantly in the realm of Web
>>>Services. Although for PHP applications as a whole typing has proven
>>>itself largely unnecessary, without it making PHP a viable and "simple"
>>>language for the creation of web services isn't feasible. How am I to
>>>create a WSDL document from a class I would like to expose as a web
>>>service? Do you honestly expect me to hand-write a WSDL document for my
>>>object and maintain that WSDL alongside code changes manually? WSDL
>>>documents were designed to be generated automatically by the
>>>architecture exposing the web service, and without any notion of typing
>>>in PHP at all there is no viable way to really do this.
>>>
>>>Also let me point out that I am *not* suggesting that PHP become a typed
>>>language. I am merely am suggesting that the ability for me to enforce
>>>structure on my objects does indeed have a real benefit without breaking
>>>backward compatibility or otherwise compromising the spirit of PHP.
>>>
>>>John
>
>Exactly!
>
>--
>PHP Internals - PHP Runtime Development Mailing List
>To unsubscribe, visit: http://www.php.net/unsub.php