[FLOW3-general] Rumours of Phoenix dropping ExtJS?

["Christian Müller (Kitsunet)"]

> Just before the RC release of TYPO3 4.6 I have not much time left to
> view videos; that will come later.
>
> NOTE: if things below sound like harsh criticism, please beware that it
> isn't meant like that. It may partly be because English is not my first
> language and partly because I'm passionate about TYPO3; I want the
> future to be bright, I want current users of TYPO3 to keep using it for
> ages and developers to find it worthwhile to continue developing for
> TYPO3. Please keep this intention in mind when reading the rest.

Don't see it like that, it is just fine. That's why we are a community, 
everything should be open for discussions. (Well basically that is how 
our decision came about, it wasn't that easy to just drop quite a lot of 
the code we had already). Please treat my answer likewise, I just try to 
give you as much insights on our decision as possible.


> A few things to consider:
>
> - during T3CON09 there were presentations about the new TYPO3v5 and also
> some sessions with ExtJS (now Sencha) developers. Everybody was
> enthusiastic about using ExtJS for the v5 UI and it was agreed that the
> ideas in the mock-ups could be realized with ExtJS. It seems like a good
> contact between ExtJS and TYPO3

Can't comment on that as I entered development from 2010, but yes that 
was how I experienced it too. Actually I think you can do almost 
everything with any framework you choose, the question is when it will 
be finished.

>
> - extension developers have to learn techniques or get used to
> frameworks. People invest time in Extbase/Fluid based on the promise
> that this is the future of extension development and that their code can
> easily be ported to FLOW3 packages. Large parts of the TYPO3v4 BE and
> system modules are converted to ExtJS and together with the previous
> point developers would expect that investing time in ExtJS will help
> them for TYPO3v5 too.

True and the Extbase / Fluid knowledge will be still valuable. Even the 
ExtJS may be useful. With our new approach (web-style app) you have much 
more freedom. In the previous development we generated everything with 
ExtJS om the client, there were no page reloads (desktop-style app) so 
every module would have to be written in ExtJS, totally adjusted to our 
API (Which might have gotten a big task to rewrite any v4 ExtJS module 
to v5). Now we generate markup on the server-side and enrich it on the 
client-side and that allows you much more freedom with your own module 
because the main menu bar is (almost) static html, so you could (maybe - 
as we haven't got the concept finished) integrate whatever you like, but 
we will stick to Sproutcore.

>
> - loading yet another framework (200KB) does matter. Even if it is
> locally cached it must still be loaded and processed in the browser with
> every page. It takes time, processing power and memory, and will thus
> slow things down. It increases the chances for conflicts and other errors
>
> - using multiple frameworks requires knowledge of all these frameworks
> for maintenance, new features and bug fixing

Right, that is why it would be nice to get rid of ExtJS, but Frameworks 
these days have become much more compatible, jQuery can run alongside 
almost anything else and Sproutcore itself needs jQuery but apart from 
that shouldn't interfere with anything you add. Let me stress again that 
before we also had ExtJS and jQuery just because Aloha needs jQuery. So 
the situation is not that different.

>
> - TYPO3v5 would be a nice demo platform for Sencha; good contacts
> between both teams would be mutually beneficial. If paid support helps
> the relationship it would be worthwhile to try and find
> budget/sponsoring and maybe get cooperation from Sencha to get premium
> support for the basic price.

Uhm, yes, but isn't that true for any other framework we use? I think 
the Sproutcore guys are just as happy to have Phoenix as demo platform.
I think the premium support just wouldn't help, we worked with ExtJS for 
over a year and we had the experience of cumbersome and bloated numerous 
times during that time. Just compare Phoenix master before Dev Days11 
with what we have done from Dev Days11 until now. You will see that we 
are almost at the same state, in some eras even more advanced. It just 
didn't seem to fit for us, but from my limited point of view I think 
that ExtJS is the right thing for v4 as the components are very nice and 
you can just add more and more of it to the backend, but as we are (as 
you say below) planning to do something totally new it just didn't seem 
to fit.

>
> - even though the code base is different (we do have Extbase as a
> backport of FLOW3 in the v4 core...) and the UI concepts are a visionary
> revolution, Phoenix will still be the next (super)major version of the
> CSM known as TYPO3. This also means that we want current users of TYPO3
> to use v5 in the future. Starting with the Transition Days a path has
> started to make a (slow) transition towards v5. There is thus a
> responsibility for the v5 team to make this transition path possible. If
> v5 is an ever moving target it is hard for the v4 team to make a
> reliable path possible.

I don't think there will be a transition of modules from v4 to v5, you 
can surely move an extbase extension to a FLOW3 package but I don't 
think that should and will hold true for the user interface, as it 
wouldn't fit Phoenix anyway, ExtJS in or not.

How do you expect this transition path to be? Would be interested to 
hear your expectations and wishes.


>
> Back to fixing some bugs...
>

:) Great, will do that too