[Typo3-dev] Quality development

[Daniel Hinderink [TYPO3]]

Hi folks,

Revived by the publication of the admin_interface extension, it has become
very clear to me again, that we have a basic problem about the development
of extensions.

When the extension system was discussed last summer, the main argument was
to split the development initiative, so that functionality could be added
and bettered by many people, instead of just Kasper and René.

While this has worked our for adding new functionality, it does not
seemingly work for improving existing functions or when creating important
functionality of greater reach, like FE-Functionality.

Let's examine this case as an example:
There has been a number of discussions about this subject on this list,
which lead to Julle, Robert and Daniel T joining forces, by exchanging
thoughts and concepts and Dt creating a new approach which was then given to
the team to evaulate. This is still under way.

At the same time, Roger Bunyan decided to create another solution, not
knowing of the discussion and it's results.

I have discussed this with Roger and while I am very unhappy with this
incident and believe he should have researched this subject, before starting
(redundant) work, he is basically accepting this mode of unorganised
contribution as a basic principle of Open Source Development.
Additionally he is saying, that TYPO3.org and our information policy in
general is permitting such situations and that we need to make sure that
people are informed about such projects.

I think he is right about the latter and that we could do better in this
department.

Regarding the mode of development and the amount of thought and research
and/or discussion required for a sustainable development of TYPO3, I think
we are facing a massive problem.
Let's face it: we have tons of useless extensions, many extensions to
extensions, instead of collaborative efforts and remaining fossils
(tt_address!) that never seem to improve.

I might call this the "*Nuke"-path and regard this an immature and
potentially dangerous path to a future system cluttered with half-baked code
and only very minor true improvement bearing the amount of time dedicated to
it. In other words: it is less productive, but also less interesting.

What we need instead is an awareness about the general path we are taking
and a more elaborate approach to collaboration.

Other Open Source developments, like the Linux Kernel do succeed in creating
teams and joint efforts simply through their mailing list.
The TYPO3 community too often fails in this respect.

There are many ways around this, from creating a company running TYPO3
(ezPublish, Zope) to limiting contribution like Linus does, or we are in the
process of doing with the rating of extensions to some degree. There could
also be a number of stages leading to the publication of an extension, one
being a public proposal.

I am looking forward to your thoughts and proposals.

Cheers,

Daniel


-- 
TYPO3 - get.content.right

Daniel Hinderink
Marketing, Press Relations, Strategy
http://www.typo3.com