[TYPO3-mvc] TYPO3-Core complete rebuild to realize modern development paradigms?

[Gabriel Kaufmann | Typoworx]

We are developers also trying to provide patches and bugfixes for 
TYPO3-Core and came to the idea, that some parts of TYPO3-Core are just 
too much got stucked in it's current code-concept.

To name only some of the main problems (as our own personal opinion):

 1. The "old" antiquated use of string-relations in TYPO3 (TCA-Type
    "Group" / "DB") saving it's relation internally as String
    ("[TABLE]_[UID]").

    There are better alternatives in times of modern programming and
    ORM-Database tools like (to name only one), Doctrine.

    More about doctrine:

    http://en.wikipedia.org/wiki/Doctrine_(PHP)
    <http://en.wikipedia.org/wiki/Doctrine_%28PHP%29>

    http://www.doctrine-project.org/


 2. Usage of the old-styled relations mentioned in 1. are more than bad
    and dirty to handle in f.e. "foreign_table_where" (f.e. TCA Select)
    clauses.

 3. Rendering of FlexForms is a nice to have feature. Nevertheless
    internally in t3lib processing it is real hard to realize a real
    compatibility
    between TCA and FlexForms (which should have nearly the same syntax
    / feature-richness) despite of that one is PHP-Array and the other
    some extended XML-Format.

 4. Extbase is a Core-Extension that still is "unstable" (even if marked
    as stable by the developers) and still misses even banal
    Database-Functions GROUP BY, FIND_IN_SET, CONCAT and "COUNT" is not
    always working as expected (in some more complex queries). All in
    all... Extbase is really hard to develop on some more complex
    database-relations and queries.

    We also tried to get in touch with the developers trying offer help
    trying to integrate a some more established database core, like
    Doctrine we already use for some other (non TYPO3) projects with
    huge enhusiasm as there it really works as expected: out of the box
    and like real "rapid development" - which we currently search
    forgiven in Extbase.


As we are working in production with TYPO3 and also realizing customer 
projects of course we sometimes come to points finding some (even really 
simple looking) features that should have been working yet - but they 
don't do as expected.

*We also noticed in the last weeks that some feature-requests and 
bug-tracks on TYPO3-Forge/Issue-Tracker have been (in our opinion) 
premature rejected or have just been ignored for some month - even if we 
provided a patch that should have been provided a working solution for 
some issues.*



For us those "tiny" issues cost us a lot of time in developing and also 
headache while trying to iron out or work-around those issues. And we 
also came to the conclusion that there might be the right time to start 
something "new" upon TYPO3 to begin a new-age for TYPO3.


*As the TYPO3 6.x tree still supports and uses...*

  * the dusty PiBase-Core (introduced with the first "old" versions of
    TYPO3)
  * Extbase becomes more and more part of Core and main-features (like
    Workspaces)
    despite of Extbase_still should be alpha-state_ (in our productive
    experience)
  * Still uses code-parts of the old-dusty "4.x" tree that should have
    been cleaned-out & consequently could have been dropped
    for a new beginning.
  * A complete clean-up would have made sense, as TYPO3 6.x already
    dropped support for some features and already has only an official
    support for a bunch of Extensions yet.

... we are thinking it may be time for a parallel, complete new 
version-tree of TYPO3 trying to introduce modern state of the art 
paradigms like MVC (even in Core/TCA). Shurely Extbase introduced some 
real nice features giving a introducion in the right direction (despire 
of their own still buggy database-/ORM).


Of course a complete rebuild of a new TYPO3 "X" may lead into something 
new that may drop support for a lot of existing Extensions - and may be 
it will lead into it's own TER-Tree and something "new" not looking 
backward for old extensions. But we are convinced that only a complete 
code clean-up and rebuild can really solve the current dillemma between 
old-code and new-code. Of couse we don't want to completly drop the 
TypoScript support that we are still convinced as a very flexible 
solution that makes TYPO3 mighty in rapid development (even for 
"non-developers" for PHP).


*We want your opinion!

Do you also have interest and ideas about that idea?

What do you think of a parallel new tree that really introduces a break 
to start something completly modernized in TYPO3?*

-- 

    Best regards

    Gabriel Kaufmann