"Cornelius's grand plan" - Merging KDElibs into Qt

[Olivier Goffart]

Hi, 

I reply to the thread.

I am a Nokia employee working on Qt, I joined Trolltech in 2007, shortly 
before te aquisition. But what I am saying here is personal and does not 
represent Nokia.

Regarding release cycles, we want to improve our releases cycle and release 
much faster. We want to have 6 or even 3 months release cycle. We tried with 
Qt 4.7 already, but we failed.  One of the reason was that Qt is becomming too 
big. This is why we want to modularize. Which mean that showstopper in webkit 
or multimedia, would not block the release of other components. This hopefully 
will help us reach the goal of short release cycles.


You might have all noticed by now that focus of Qt developer has changed 
towards mobile stuff, leaving desktop behind.
And this is where the contribution model and open gouvernance comes into 
account. 
We have always wanted to open the developement as much as possible. We have 
already acheived a lot, and want to continue. This is unfortunatelly a slow 
process, but we are getting there.
The open gouvernance would allow KDE or other people to take over the 
maintainance of Qt classes that are relevant for KDE. This include obviously 
push access and everything.   But we are not willing to sacrifice our quality 
standard, which are different from current KDE standard. (This can be 
understood as coding for a library is different than coding an application)


Regarding the licence agreement, I cannot comment on that. But I will just 
point out that contributors keep copyright on their code, and that many other 
free software projects (notably the ones from the FSF)  require to give such 
rights to an entity.
But I can understand that in its current form, it gives an unfair advantage to 
Nokia.


So back in 2005-2006, in the time we ported KDE to Qt4, one of the goal was 
already to reduce kdelibs size by using as much Qt classes as possible, and 
try to remove the KXxx if there was a QXxxx.  So this is not a new goal.
We managed to make it work for many classes, but still many classes could not 
disapear because the KDE equivalent provided more features or was way  more 
advanced. I blame Trolltech on that, they should have looked at what KDElibs3 
was providing on top of Qt to be able to provide the same level of features.
Now Qt pays the price for that, because integrating a pure Qt application in 
KDE is much harder.
And things have not changed. This is a problem of mentality. I feel like most  
Qt dev just do not care about KDELibs or other 3rd party library built with Qt 
and what they could bring.
But having more KDE people involved into technical decisions regarding Qt will 
definitively help, I want that to happen! 

-- 
Olivier



Le Sunday 31 October 2010, Mark Kretschmann a écrit :
> Hey all,
> 
> after reading the whole thread that started with Chani's mail ("why
> kdelibs?"), I think the noise level has become a bit too much there.
> Cornelius had proposed this rather daring idea:
> 
> http://lists.kde.org/?l=kde-core-devel&m=128842761708404&w=2
> 
> 
> It's a very controversial idea. However, I think it is so refreshing
> that it deserves some more thought. Personally, I think the idea is
> great, if we can overcome some of the obvious road blocks. I would
> love to read some honest and direct thoughts from you guys.
> 
> 
> What do you think about it?