On 24/03/2014 10:52, Johannes Schlüter wrote:
On Mon, 2014-03-24 at 10:42 +0100, Pierre Joye wrote:
Agreed on the contributions process, however:
Ivan is talking about specifications, an implementation follows a
given specification, not the other way 'round. A well written
specification greatly simplify documentations, implementation(s) and
testing, to name only a few advantages.
Yes, that's what we do in RFCs. I won't stop anybody from consolidating
a base specification, but as everything in this volunteer driven project
this depends from people actually doing this, not from meta discussions.
This is not a meta discussion at all. This is a real discussion because the usage of Computer Science are evolving. Most of languages have a specification: Javascript (with ECMAScript), Python, C, OCaml etc., with different interpreters or compilers (respectively SpiderMonkey or V8, CPython or PyPy, gcc or LLVM…).
Also as long as PHP is the PHP reference implementation all things have
to be implemented there to be part of the language.
And here is the first conflict. While I deeply agree with you, this may not be how we should work. We should at first create the specification and then secondly provide an implementation that respects this specification. It would be great to see internal@ to be the first interpreter to implement the standard, but this will not be always the case. Another project could propose a new feature along with an implementation in their own interpreter; when the feature will be accepted, then we will implement it in the “historical” interpreter.
I think we have to split the language and the implementation, while keeping php-src as the default and (so far) the prime implementation.
That being said, I find very disturbing the total lack of interest
from historical (from a time pov, not activity) core developers in the
next major version. Most drastic changes around PHP happen outside
PHP, many features our users are looking for are implemented outside
the core while we keep arguing about the needs of these features. This
is not a good thing.
In my opinion it is a good thing that PHP is powerful enough to enable
frameworks etc. to innovate around the core language. The core language
has to be a reliable and stable.
Yup, +1. But I think that Pierre was referring to PHP6 features and the weak interest and responses he collected (but I don't want to speak for him).