Hi guys,

I have two questions:

1) Is it legal to fork the LGPL/GPL parts of Qt? Could Nokia do anything about this if the fork's name is not Qt ?
(I posted the question in this forum, because I read about the KDE Free Qt foundation: http://www.kde.org/community/whatiskde/ ... dation.php)

2) What do you think about the several direction changes in Qt5, the general movement from the desktop to the mobile phone area.
Wouldn't it be much nicer if there's a Qt fork, which deals with the usage of QtGui libraries from C++ only? All other stuff for mobile phones can stay at Nokia, they're the experts for mobile phone stuff. But why not create an open gui c++ only library which runs on Windows, Linux and BSD just without the mobile stuff and - if Nokia continues to abandon the C++ libs and uses its development resources practically only on the mobile site, prefering to use C++ as a backend only - I suggest to fork it!

Qt 4.8.1 would be a good point to start from. I think that there are many interested developers, who share their aversion against QML and mobile phone direction of Nokias development plans in the qt forums or qt-labs blog comments, who would therefore be interested in such a fork.

I also do not think that forking is generally a bad thing to do. If we look at BSDs, there are many positive sides: For example, FreeBSD benefits from DragonFly BSDs improvements, bug-fixes, etc. too! Same could be for Qt and a new open source c++-only qt fork.

Best regards

Qt upstream still has many interested parties aside from Nokia itself which are dependent upon the Desktop version of Qt.

There is no reason to fork it at all, and the new Community involvement process allows these interested parties, including KDE to be involved in maintaining the parts they need.

Ok, so if they drop their engagement to the Desktop version, you or the other interested parties - who are contributing already - take over?

But just hypothetically: Would forking be legally possible?

I do not see a reason for a fork here. If you want to contribute patches related to classes used in Plasma Desktop (or whatever other software not used in mobile) you can already, no fork needed.

What do you have against QML? It makes creating GUI code a lot easier (unless you need keyboard handling), smaller (less lines of code), simple to understand, we do not have to worry about binary compatibility since it is interpreted and you can always create a C++ plugin to handle things that QML is not suitable for. Plasma is already in the process of replacing the current QGraphicsWidget plasmoids to QML too.

This thought about forking was just hypothetically. In fact, I was disturbed about Qt5. And if it turns out to be like this, a fork might indeed be necessary, but until then, I will of course try to engage in the main project, send patches, etc.
But I assume, we could make the whole Qt framework much smaller and cleaner without the mobile phone part, which does not really belong to it (in my humble opinion). Qt became famous with the C++ desktop libs, besides, Nokias presence on the mobile phone market is a disaster, if you compare it to iOS and Android. So, is there really a future for QML, which clearly was targeted at that market? And if you watch the commits in the Qt master repository, most of them are about mobile phone stuff or their C++ backends, QML/QtScript, etc. Not many changes/improvements in the classic stuff, besides bug fixes.

To QML:
Yes, this may be true for "mobile phone"-like interfaces, but in the desktop world, it seems to be only another level of abstraction, a new language (not really javascript but very similar) for things you can achieve with C++ too. It's rather clear to me that this push towards QML and mobile phone stuff comes from Nokia, because their main product line consists of mobile phones.
I do not understand why it should become more simple, you still have to know C++, even if most GUI stuff is handled via QML.


So nobody of you was ever upset about the decreasing interest (and decreasing use of Nokias resources/developers) on Desktop Qt ? Sure it isn't abandoned (yet), but it definitely deserves more attention from the Core developers.

The way Qt 5 is being constructed, it allows for certain modules of Qt to be excluded if they are not needed. Given the "mobile phone" parts are not tied in heavily to the other parts, then they could certainly be excluded.

However, the "mobile phone" part being referred to is likely QML - which likely will provide some very interesting scenarios for future development as Lamarque noted above.