[TYPO3-ect] 5.0 roadmap - continuation or new branch

[Elmar Hinz]

Michael Scharkow wrote:
> I never proposed starting from scratch, but I argued that a lot of the
> components in TYPO3 core are basically too complex to develop or even
> maintain, like TCEMAIN, so compared to refactoring old stuff a rewrite
> seems even less work. However, this does not at all mean that we start
> with a blank file because a lot of components can very well be used, so
> the whole development is incremental rather than from scratch.
> 

Hi Michael,

it's allways difficult to discribe the proposals of someone else precisely.
Please forgive me the variations. I hope in general people can catch the idea.

> According to Robert's (not my) proposal there are even three teams: One
> maintaining 4.x for a time, one working on 4.5 and one for 5.0 .

Hmm, do we have the manpower for 3 teams?

b) 4.x should be maintained by those who are interested in it and have it in
> production, 

This seems an important point to me. It is the interest and task of those,
that have it in production. How do you porpose to involve them? How do you
propose to organize this team?


>> Old extensions would continue to run in the 4.x branch, but mayor
>> alterations would be necessary to port them to in 5.x branch.
> 
> I don't see why this should be the case. The core-extension connection
> is rather thin and I am sure that *backward* compatibility is no
> problem. But you seemed to be worried about forward compat, so that new
> stuff runs on old TYPO3 or PHP4. This is impossible without sacrificing
> all the advantages of the new platform.
> 

The current extension API is not very beatiful. :-) We should consider
carefully, if we really want to implement the old API into a 5.x candidate
or if we take the change to design a new one. There are pro and contra.

If we really wan't to build a "clean" 5.x, we also need to discuss such
basic things like the db model. A change could have impacts on many
extensions. I.e. the plugin point of plugins could be alterd then.

I am not worried about compatibility myself. I rather wanted to point out
advantages and disadvantages of the 2 alternatives.

If we do this split, I postulate, that the 4.x branch is maintaned for at
least 1 year after 5.x has become stable. The corridor to port extension
from a stable 4.x to a stable 5.x branch (and vise versa) should be broad
enough.


Regards

Elmar