[TYPO3-dev] TER, typo3.org, everything went FUBAR!!! :))

[Daniel Hinderink [TYPO3]]

Hi,

Thanks for your concise description of more reasons for the current
situation, again proving that there is no shortage of knowledge, but one of
action. However, your description does not change my view about what I think
the initial factors for this development are, it just adds to the
description of the chain of events resulting from it.

Unfortunately, the output of this discussion in terms of problems solved is
0, if it does not motivate anyone to join in projects like TER2 in time.

How about you Dimitri? All the time wasted on this list put into development
would be nice break for everybody.

Cheers,

Daniel


in article mailman.7722.1141605530.3122.typo3-dev at lists.netfielders.de,
Dimitri Tarassenko at mitka at mitka.us wrote on 06.03.2006 1:38 Uhr:

> On 3/5/06, Daniel Hinderink [TYPO3] <daniel at typo3.org> wrote:

> What my mail
> was about is that this project is as good as we, the community,
> make it. In
> this case, this community has neglected beta testing and very
> few people got
> on the project to help create TER2 in the first place. Now we
> have problems
> and a bottleneck - surprise?

Let me offer you a couple of more reasons to why
> it happened - and I
am taking these guesses based on Robert's description of a
> problem,
correct me if I am wrong:

1. Someone very enthusiastic decided it is
> a good idea to run a huge
production environment that is typo3.org on the
> "latest and greatest"
version of PHP, self-built, on top of that.

2. Robert's
> development and testing environment was different than the
one TER2 was
> deployed into. (PHP version and platform were different).

3. Instead of using
> PHP-based and much better tested SOAP
implementations, the decision was made
> to go with PHP SOAP extension
in PHP5 (this assumption is based on Robert's
> comment that he was
debugging PHP SOAP ext code). There may be valid reasons
> for that that
are not obvious, but ceteris paribus, I'll take code that's 3
> years
old over the code that's 1 year old any time.

4. Some bug (apparently
> in PHP5 SOAP ext.) had screwed stuff up.
Instead of rolling back PHP on the
> server to the version he used in
development (not quite sure if he has control
> over this), Robert
decides to help the guys at Zend and starts debugging PHP
> SOAP code.
Been there, done that, guilty as well. I spent three days
> once
debugging PHP interface for FDF Toolkit, while now, looking back,
> I
could've solved the problem in a couple of hours with XFDF instead and
let
> much more qualified at this task people at RedHat or Zend deal
with PHP
> issues.

With all that said, there is a chance beta-testing could help
> find
this problem earlier, _if_ it was carried out in the same environment
as
> the production server. Whether that was the case or not I don't
know, you tell
> me, but I really doubt it - Robert would have found it
himself very
> quickly.

So, when you say "if you don't participate in beta-testing,
> don't
expect stuff to work and don't complain", I think it's a
> bit
far-fetched.

That's as far as the SOAP problem goes. Now, as far as
> documentation
conversion and broken links are concerned - do you really need
> "beta
testing" to see it's broken?

> As long as you (meaning anyone) expect
> yourself to make it high quality, I
> think there is a chance that it will
> happen. For those that are here to
> socialize on the mailing lists, they can
> only hope someone else does it for
> them. What they certainly have no right
> to, is to complain.

So, I take it, you are limiting the access to bug tracker
> to the
developers only? This way, you get rid of all these
> meddling,
complaining socialites once and for all.

When people complain, it's
> typically not because they hate you,
Robert, Kasper, TYPO3 or anyone else.
> It's because they see something
that doesn't work, is really a problem for
> them, and they try to get
your attention to fixing it. The squeaky wheel gets
> the grease. What
you are proposing is earmuffs for the driver.

I think Robert
> is doing a great job - the only problem it is put into
production a lot sooner
> that it should have been, and there is some
disconnect in
> testing-staging-deployment process.

--
Dimitri Tarassenko