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

[Andras Mantia]

On Sunday 31 October 2010, Mark Kretschmann wrote:
> 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
[..]
> What do you think about it?

I will ruin the show, and say: do we want this? What is the goal for KDE 
(and not KDE eV, for which the goals are written down). What do we want 
with KDE itself? I will say what I want: the best desktop available, 
that is a pleasure to use and pleasure to write applications for it. And 
I'm egoist enough to say, that I also want the fame part: that it is 
recognized that this product was created by people who believe in open 
source and this people are those who form the KDE project.

 Is it the best desktop available? For me yes, but it is not, in 
general. Unfortunately it still has quite some bugs, in many areas. I 
always try to believe how nice and easy to use is, and see that I end up 
apologizing for this broken feature or that, or fixing myself the 
desktop of others, who are beginners in KDE. So that is clearly work to 
do. And where we need to work nowadays is mostly the applications. Not 
only (e.g a notable exception being the printing system and another 
famous bug I don't want to mention directly), but in many case. They 
want a desktop that doesn't crash, a browser that just works, an email 
application that is easy to use, and so on.

 Is it a pleasure to write apps for it? For me, yes, it is. In the "why 
kdelibs" thread, lots of the nice technologies were listed. For 
newcomers, again, it isn't perfect. And I don't say newcomers, who are 
interested in one or other part of KDE libraries, but who are interested 
in KDE as a whole. This is mostly a documentation issue, though. The 
library code itself I find to be good quality, with much less bugs than 
in the applications (now somebody can give me bugzilla kdelibs vs. apps 
statisctics to prove me I'm wrong...I talk about my daily experience).
The monolithic approach and being an extra lib on top of Qt might also 
scare some developers. The question is, how much do we sacrifice to get 
those developers. Do we break SC and BC again to try to do it "right", 
and piss off all the current app developers, who need to port they 
lovely project again. And do the same with the users, as there won't be 
regression free porting. I'm not against reducing inter-module 
dependencies, or making it easy to check out part of a library and 
build/install only that, but I'm against doing a full library 
restructuring which requires the application developers to port their 
application to a new version. Remember, we still did not fully port the 
applications to KDE4 technologies! And I bet there are still quite some 
Qt3 and KDE3 support module usage in the main kde modules themselves 
(eg. korganizer was cleaned up only recently).
 Instead of doing it, we should provide good code examples, good 
tutorials and a good tool for developing. Like Qt Creator is for Qt 
projects. Be it an extension of Creator, or even better a good KDevelop 
(which has the same problems as of now as I said before: it is buggy and 
feels unfinished).

 And then something that is about fame and recognition: we tried to 
build a brand, we try to show that open source is innovative and can 
produce cool technologies. I'd not like to see that our work disappears, 
by being merged into something (Qt). If it happens we will sadly see 
what we have now with webkit and also phonon. Companies have no idea 
they are KDE technologies. I have no problem with Qt itself, it is a 
very good library and the base of KDE. I'm glad Nokia made it available 
under free licenses. I'm glad they are more open than ever. But it is 
still a product of a company, it is not the product of a community. Even 
if former and current KDE people work there. And this is related to 
licensing: do you want to give your code to a company to do whatever 
they want with it? To market it as their product? I have no problem if 
somebody makes money based on KDE, but I'd like to see that the credit 
is also given to KDE. 
 
So what do we need? I think we need to work on three areas:
- advertize the good things we have in the libraries (marketing 
material, tutorials, blog and forum posts also in non-KDE related 
websites, even books)
- make sure that people can actually easily use what we advertize 
(tutorials, API documentation, development tools)
- bugfix our applications as much as we can, so end users enjoy the 
power of our platform

I know, people cannot be forced to do this or that in an open source 
project, and we shouldn't do, still I think the above should be the 
goals of the project and the steps that needs to be done in order to 
achieve the goals.

Andras

 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 190 bytes
Desc: This is a digitally signed message part.
URL: <http://mail.kde.org/pipermail/kde-core-devel/attachments/20101101/75df0042/attachment.sig>