Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:73388
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: error (pb1.pair.com: domain schlueters.de from 217.114.215.10 cause and error)
To: Larry Garfield <larry@garfieldtech.com>
Cc: internals@lists.php.net
In-Reply-To: <5330661B.80202@garfieldtech.com>
References: <532FF7B9.5040700@hoa-project.net>
	 <DD864C7C-6497-411A-9815-36D7662B2846@ajf.me>
	 <5330582B.1020604@hoa-project.net>  <5330661B.80202@garfieldtech.com>
Content-Type: text/plain; charset="UTF-8"
Date: Mon, 24 Mar 2014 20:10:34 +0100
Message-ID: <1395688234.9365.572.camel@guybrush>
Mime-Version: 1.0
Content-Transfer-Encoding: 7bit
Subject: Re: [PHP-DEV] PHP Specification
From: johannes@schlueters.de (Johannes =?ISO-8859-1?Q?Schl=FCter?=)

On Mon, 2014-03-24 at 12:06 -0500, Larry Garfield wrote:
> 
> It's not just "well someone should write it", because it's not just a 
> question of writing it.  "Just write it" is a documentation effort,
> not a specification effort.  A specification effort would entail a
> cultural shift that new RFCs are patches against the specification
> document, NOT against php-src.  A patch to php-src would come second,
> and would be bound by the specification, not vice versa.  That is,
> this isn't just a matter of "somebody should do something", it's "we
> should collectively change our process."
> 
> The proposal on the table isn't "we should write a spec version of
> what php-src does today".  It's "we should change the approach to
> evolving PHP to be specification-centric, not
> reference-implementation-centric."

PHP works a lot by pushing contributors to change their habits over
time: the RFC process started slowly and evolved and will continue to do
that, it will more and more look at a specification level. In that sense
there is the need for somebody to start writing it. This won't happen if
nobody sits down, creates an initial version and pushes for it.

That said: Even a specification driven process will need feedback from
some implementation. (C++98 was never fully implemented by anybody as it
was basically impossible, to do, some (all?) these issues where fixed in
C++11 which had two implementations close to the standardization
process, with C++14 the two are even faster than standardization; is
there any fully standard conforming Perl 6 implementation, yet?)

So yes, we need somebody who *pushes by doing* till we find a good
balance. it won't happen by itself. 

johannes