[TYPO3-dev] Little summarize

[Sven Wilhelm]

Hi,

thanks for your open discussion until yet. I'll try to summarize a
little bit of the current discussion. A possible new newsgroup is
waiting for Michaels respect Kaspers Go. So only few discussion yet :)


At the moment the most discussion focus the question if or not the
smarty template engine should be the new engine for TYPO3, how it could
be integrated and corporate with Typoscript. Discussion without any
direction.

Another part of the discussion was wether it's better to use standards
like XML or keep the current way TYPO3 is working with arrays.

Specially JoH told to keep an eye on existing users and their
investments, be it budget or time to get their knowledge to the current
level they stay. Summarizing his statements the core 5.0 should be
refactored, what is a difference to rebuild it.


More interesting facts could be what current problems are and how they
can be managed.


>From a designer/xhtml builders point of view
---------------------------------------------
Building XHTML/CSS is a job currently seperated at most from TYPO3.
These guys can work with their native tools, test the work outside and
try to integrate in the last step (ok I'm not a designer, so please
correct me if I'm wrong).
The current style having simple markers in the template is practical, it
works.
What could be improved there?
TYPO3 has it's own template system, means it must be developed by the
project itself, second point, for complex stuff it's not easy to use it.
Possible solution: Use a standard system with the knowledge of several
implementations that works as similar as possible to the current system
(designer view).
Therefore I just focused on PHPTAL.
Best solution: Have a template API (as some told here).


>From the developers point of view
----------------------------------
Developing TYPO3 is currently extrem hard as it is written extrem
procedural, even if you see "classes" and "methods". Just build a
sequence diagram of something you're interested in.
Another point is the extrem usage of $GLOBALS, breaking all layered
architecture as you can access most data on any layer!
The amount of data stored in arrays is another point. Ok, if you say PHP
is optimized to handle array data. But how much data I necessary to
answer a simple request?

Today at least every developers tries to use a good IDE, be it Eclipse
or Zend or whatever (maybe some use als vim :)
These IDEs support you to build good code as long as they are ABLE to
help you. Supporting the developer is possible as long as he uses some
standards.
For example a unit test is hard/impossible to write if you cannot expect
the same peace of data from a method as true and the rest of cases as
false. But this kind of thinking should be implemented in the brains
now.

At least new developers only join if developing isn't masochism (or it's
paid).


>From the administrators/providers/agencies point of view
----------------------------------------------------------
Be it a single person with one or many installations or a company
offering services to thousands of customers, the wish is (let's tell me
something other :) to distribute instances with the smallest amount of
work necessary. If you can offer your customer additional services like
monitoring and systems management, perfect.

What's necessary for that? Having standard interfaces like soap for
remote management/monitoring. To implement this you can hack around and
get things extended or working on a clean core where an api provides you
(think a little bit about protocol dispatcher, a protocol is nothing
more than a standard defined access method (with the individual
inspirations from companies :))


>From the editors point of view
--------------------------------
The best system is a 100% intuitive systems. There are many scientists
out there researching how user interfaces should build for a maximum of
usability. As a practical examples (please not flame war on that) GNOME
and OS X use a complete different desktop style as KDE and Microsoft.
The first two simplify the desktop, the last two bloat it..

Many editors are not power-users, means they edit content from time to
time. If the usability is bad, the support for such users is heavy. And
even imagine, also a simple desktop is sometimes to complex for such a
user!!!


I think the best way to build a wish list for a new core is to think
from outer to inner, killing all holy cows first (you can reanimate them
later :)

Fresh thinking without any barriers is the most stimulant stuff (for a
time).

Cheers Sven

Ps: Asking someone: "Ok, so why you use it then"
    Possible answers:
    -> Because I invested time
    -> Because, even if it lacks on the current limitations, other
       solutions are not better.

I like TYPO3, see my feedback as input to improve it!