Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:12349
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Date: Wed, 25 Aug 2004 17:32:44 -0700
Message-ID: <C043E4034185BB40B01C76980AC639C066F1D0@cacexc04.americas.cpqcorp.net>
Thread-Topic: Unfulfilled promises... forever experimental extensions... all over again
Thread-Index: AcSLBDQj6d+RO+5FQFyQXCyLg+0Ziw==
To: <internals@lists.php.net>
Subject: Unfulfilled promises... forever experimental extensions... all over again
From: roshan.naik@hp.com ("Naik, Roshan")


Every once in a while somebody has to speak out about the state=20
of the union and bear the brunt of some anger... with a hope
that things will change for the better.

2 years back I started out asking a question being
concerned by the frustration caused to many users (and vendors=20
alike). Frustration due to many of the useful extensions forever=20
being in experimental mode for the longest time. I asked ...

  " And exactly how long is it expected to continue to stay  =
EXPERIMENTAL ?"

The response was a predictable one, expressing indifference...

  "Between 15 days and 15 months"

http://marc.theaimsgroup.com/?l=3Dphp-dev&m=3D103159872214635&w=3D2


So I decided to analyze CVS repository and posted the conclusions...
http://marc.theaimsgroup.com/?l=3Dphp-dev&m=3D103161600832402&w=3D2

It included an analysis of how long each of the extensions had been
in that mode.=20

To point out a few:
    sockets, openssl, crack, domxml

 *had been* in that mode for 2 years ( i.e.. more than 15months). =20
And today they are all still experimental. So that's their=20
4th (experimental) birthday.=20


Now just for the kicks lets take the case of Sockets extension and =
analyze its
history.

There was a msg on 21st feb 2002 on this mailing list.=20
Here we see some kind of encouraging remarks about achieving=20
stabilization by 4.2.X...
http://marc.theaimsgroup.com/?l=3Dphp-dev&m=3D101424999803151&w=3D2

<quote>
   Timeline & Stabilization
   =
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=


   I would like to fix all of this by 4.2.X. I propose that we then mark
   the extension as stable, and freeze the API.=20
</quote>


on Sept 6 2002, a frustrated user asked..
http://marc.theaimsgroup.com/?l=3Dphp-dev&m=3D103134854829058&w=3D2

<quote>
   Do the maintainers of the PHP sockets extension ever intend to get =
their act
   together and cement the API, or are people who need this =
functionality going
   to be forced to rewrite their code every time there's a new =
point-release of
  PHP?
</quote>


and the response was...
http://marc.theaimsgroup.com/?l=3Dphp-dev&m=3D103135145531682&w=3D2

<quote>
  Which part of "EXPERIMENTAL" in the docs at [snipped url] don't=20
  you understand?

  If you're so desparate to have things cemented, write a patch,
  post it here and we'll commit it.
  If you don't have the skills, you might consider offering those that
  do some positive incentive to do it for you.
</quote>

Basically...do it yourself or buzz-off attitude. And this is not the =
only
occurrence of such backlash.

After some folks stepped up the pressure, very encouraging comment on=20
sept 12  2002 that it will indeed make into production soon....
http://marc.theaimsgroup.com/?l=3Dphp-dev&m=3D103179363830424&w=3D2

<quote>
   Most of the other work needed to mark the extension stable has been
   done, all that remains is that I have some small win32 work left to =
do.
</quote>

sockets, openssl, domxml are very useful extensions. Will these (and all =
the others)=20
ever get out of experimental ? doesn't look like. How about the shiny=20
new SimpleXML or sql-lite?  If history is in any indicator...then I =
guess they will
suffer the same fate. And likely that their users will also be slapped =
in the face.
domxml has already had a super long experimental career.
And now all this new stuff has gone into it. So I guess that makes it
even less likely to make it out of experimental. Please don't think
I am picking out these due to some vengeance for the people behind it.
I am giving examples to see where we have been, where we are=20
and are we really going anywhere.

If these are experimental then why include them in the propaganda ?=20
Just to shoo-off the people for whom all this was created in the
1st place ?

Like the bank telling you that they will give you a "free" checking =
account.
Then when you complain your money isnt totalling up to ur deposits,
they shoot back at you saying ... "checking accounts are experimental=20
and it was mentioned in the brochure"


Today I see PHP5 with a set of promises that are congruent with the=20
earlier PHP4 release.

From the php4 release I can recall some of the propaganda...

 - redesigned scripting engine (much faster, much better, much ...)
 - Under UNIX features a much smarter and generalized build process =
(smarter, nicer )
 - enhanced object oriented programming - (now we've got it right!)
 - new extensions with cool new features
    +  Java support (still experimental)
    +  XSL          (still experimental)
    +  SOAP         (still experimental)
    + ...

The same mantra being repeated is a sign that we didn't get it right=20
the last time we chanted that mantra....and by inductive inference this=20
effort is also questionable.

But the first few are all right to promote as they actually made it into =

the product, got used and are properly supported. But blowing the horn
about new functionality included in experimental extensions should be
deplorable and considered dishonest practice.=20

Hopefully belts will be tightened up around the extensions and the
responsibility taken up by their authors.

-- Roshan=20