[TYPO3-dev] RFC: Change roadmap for 4.5 and 5.0

[Daniel Hinderink [TYPO3]]

Hi Michael,

A lot has been said about this, but I would still like to return to this
original email to add another perspective that needs at least as much
attention when planning the roadmap as the technical tasks at hand. That is
the towering challenge of having to manage the development process.

The complexity of that challenge is not only driven by the amount or
complexity of the code to be written, tested, debugged and doucmented, but
also by the amount of programmers and entailing styles, beliefs etc. to
synchronize. The likeliness of such a project failing completely are
considerable, if not high.

After all it is already very hard today to work together under the
circumstances of distributed development. When the common ground (e.g.
Backwards compatibility with a system people know pretty well) is no longer
available, the different strands of development might never be possible to
match!

So, the #1 organizational reason to work in steps like planned, is to 1.
(Version 4.0) build knowledge and problem awareness about working together
within the known "universe". 4.5 will attempt to reach out into changing
(and thereby modularizing) parts of TYPO3, that lend themselves to immediate
feedback-loops. I am sure much of the work done for 4.5 will help to think
about a new core from the pov of ease of use (including refactored code) and
usability (on all levels, including templating and such).

If the developers taking part in 4.5 development grow together to work as a
team, learn to define and adhere to self-defined principles and standards,
if all that is working out when creating and rolling out 4.5, only then I
think it is a sane option to tackle the challenge of 5.0.

Creating 5.0 is at least as much a challenge on a personal level for the
developers in this community, as it is hopefully a technical masterpiece.
However the latter will never be possible without adapting to rules of
teamwork far better than this has been the case up until now.

That being said, after browsing through this thread, I am happy to find that
much enthusiasm to start working on 5.0!

I am even more optimistic now that it is actually possible :-)

Happy Easter!

Daniel



in article mailman.1.1144665083.31122.typo3-dev at lists.netfielders.de,
Michael Scharkow at mscharkow at gmx.net wrote on 10.04.2006 12:31 Uhr:

> Dear *,
> 
> as the latest and greatest version of TYPO3 is out and 'almost feature
> complete' to cite Kasper himself, I would like to spend a second or two
> on the roadmap for the next major milestones.
> 
> In short, I'd argue that we should drop the 4.5 milestone and move
> straight to 5.0 while still improving the UI and usability of TYPO3 on
> the way. There are some reasons for this:
> 
> 1. If we change the TYPO3 core for UI enhancements (and most of that is
> hardcoded in the code), we risk losing a lot of that work again when 5.0
> is out and everything will be refactored, and most stuff even rewritten.
> 
> 2. Usability enhancements are very hard to implement if you're *always*
> faced with the holy cow of backwards compatibility. How are we supposed
> to make TYPO3 easier to use and still make everything look as if you're
> still in 3.5? The backwards-compat forces us to write tons of glue code,
> wrappers on wrappers, update-wizards and makes the code maintainance
> overly hard.
> 
> 3. UI consistency means that you have to be *opinionated* about best
> practices, and not encouraging 20 different ways of doing stuff. TYPO3
> has until now been very unopinionated because of backwards-compat: We
> still ship crappy invalid static templates *although* we know better, we
> still support HTML4 doctypes although there is absolutely no reason not
> to support XHTML only (except the said legacy arguments), which leads to
> incredible amounts of code (and developer efforts, and support mails)
> being wasted.
> 
> 4. Being tied to PHP4 sucks, and makes real elegant solutions hard to
> implement. It also leads to an inflation of
> if (function_exists(foo)){
> foo()
> } else {
> fake_self_implemented_foo()
> }
> 
> TYPO3 does its own SQL escaping, output compression, etc. only because
> we have to make sure that it runs on Grandma's VAXstation with PHP3. If
> core devs were allowed to rely on PHP5 features, we could produce more
> elegant and maintainable solutions.
> Given that TYPO3 is already a memory hog and therefore does not run in
> tiny shared hosting environments anyway, why not make it PHP5 for the
> next version?
> 
> So again: Continue the 4.x line with bug fixes only, and start TYPO3 5.0
> now, with some snapshots on the way, like the ZOPE people did with X3.
> 
> Cheers,
> Michael