I worked my way through the whole thread, finally, and I'd like to add my own 0.02 EUR.
1. Some of you are
tired of following KDE's development from version 2 to 3, then 4, and now 5. While I share your opinion that a stable desktop is a value on its own, as it is an important basis for application development, we should bear in mind that
KDE is based on Qt. So every time a Qt version is dropped by their maintainers a project like KDE has to decide to
follow the development of the underlying toolkit or to stick with the old one. The latter would imply that there would either be no patches for security issues and bugs, or that the KDE developers would have to provide the fixes themselves. So either they would put their users on unknown risks or put effort into old technology.
Of course, the switch from KDE 3 to 4 should have been a lot smoother, and the first releases of KDE 4 were immature, to put it friendly. But looking at KDE 4 now the decision to follow the development of Qt was definitely a good one.
And as has been mentioned,
Xfce is now more or less
forced to make the transition to Gtk3, although quite a few of the developers heartily hate it, just because Gtk2 will be abandoned sooner or later by its developers.
2. KDE 5 in its infancy wasn't as good as it should have been, too, but it caused a lot less headaches than KDE 4 at this stage. Now,
KDE 5 is quite usable. I just hope that more applications are migrated, soon. In particular, I would like to see Konqueror with a Blink based rendering engine, finally. Why? Well, try the latest version of Qupzilla on Qt 5, and you'll know what I mean.
3. There's a reason, why
KDE has always been more popular in Europe than in North America. The reason is that KDE is the desktop environment with
the most flexible approach to internationalization and localization. E.g., with KDE it is easily possible to have the complete desktop in your native language, without changing the system locale. This is a big advantage for many users (especially system administrators and software developers), as this allows them to get error messages from command line tools in English and post these error messages in forums to get help. With any other language the chance to find help is reduced significantly, as the global
lingua franca in science and IT is English, nowadays. At the same time the desktop and all GUI applications appear in their native language making the desktop a friendly experience for them.
4. Regarding
modularization and Gnome: No, Gnome is not modular. It's various parts are relatively independent of each other, but a lack of integration is no sign of modularization.
5. Regarding
modularization and KDE: Yes, KDE is more modular than Gnome is, but something went wrong underway... In the past one advantage of KDE used to be that the whole desktop including the KDE applications made up a release. It is quite understandable that the KDE developers had a nightmare of integrating everything with each other before a release, so that they wanted to change this. However, they created a mess, and now try to get things straight, again, step by step. Their initial conclusion was obviously to put the burden of integration work on to the backs of the distributors, but this didn't really solve the problem, of course.
6. Akonadi, Nepomuk, Strigi, Baloo (or whatever it is called now) could serve as a
reference pattern in software architecture. However, the
dependencies of KMail on Akonadi, for instance, hasn't made KMail any better. It is still one of the best E-Mail programs for Linux, but it was before Akonadi, already. Unfortunately, this multi-layered architecture is complex and sometimes suffers from
synchronisation problems between the components. Also,
Akonadi occasionally
sacrifices some of the above mentioned
advantages of KDE. E.g., if you run KMail before setting your desktop locale, Akonadi will create resources with English names. In KMail they will show up as folders named "Inbox" and "Sent", when for instance for a German user you want them to be "Posteingang" and "Gesendet". Final complaint: The complexity makes it less obvious where user data are actually stored. Which folders and files should be included in a backup? In KDE this isn't very obvious.
7.
KDE vs. Qt. Sometimes I wonder why KDE is so involved, and what it actually adds to pure and genuine Qt. As I do like Qt a lot more than Gtk I always wonder, why no lightweight Qt based desktop is available. LXQt doesn't seem to make fast progress, and lacks the capabilities regarding localization mentioned above. Which answers my questions, partially.
8. As has been said here, despite the slightly arrogant reply to Eric's post, the
KDE developers seem to have learned a lot from the KDE 4 disaster. I am confident that they are going to fix the most serious problems, like they did in the past. And it's good that they seem to listen to people like Eric. I guess, after thinking twice, they noticed that
Eric (and Slackware as a whole) are great advocates for their work, and not criticizing them arbitrarily. So
mutual support should be the
modus operandi.
As a side note, I am not surprised that an OpenSUSE developer supported Eric's point of view. I have said in many times, that I like OpenSUSE almost as much as I like Slackware, and this has to do with the people working on that distro. Like the Slackware devs they are friendly, helpful, smart and skilled persons with a good sense of humor, and like the Slackware devs they are down to earth, but have a vision and imagination. You won't find a character like LP among them...
9.
Thanks Eric! I think your post has egged them to re-think their way of collaboration with distributors, which is good!
gargamel