[The_Colonel]

KDE4 is a sad story. It ruined the reputation of KDE for a very long time. I loved KDE 3.5, but then came back to KDE only in 2021.

It's mostly just a problem of communication. KDE 4.0 should not have been marked as stable and should not been used by distros. If they baked KDE4 for two more years and on the side maintained/developed 3.5, the transition might be much smoother.

[adrian_b]

While there were some bugs in early KDE 4, those were not the main problem.

No amount of baking could have saved it.

The main problem was caused by the completely different purposes of the new developers, who have removed all the outstanding customization features of KDE 3.5.

For me, indeed KDE 3.5 has been the best graphic desktop that I have ever seen. Neither before, nor after and neither on Apple or Windows have I encountered anything as good.

The main reason for this was that KDE 3.5 allowed extreme customization, so you could make your own graphic desktop that did not resemble at all the default desktop.

After the shock of experiencing the garbage KDE 4, even if I had waited a half of year before making the transition, with the hope that any major bugs would be solved, I have reverted to KDE 3.5 for a few years, until it had become much to painful to make upgrades in such a way as to not damage it.

Then I have switched to XFCE, which does not provide as much as the old KDE 3.5, but at least it does not get in the way of your work with undesirable and hard to remove features. Moreover, any useful KDE applications, such as Kate, work perfectly fine on XFCE, together with any useful Gnome applications.

The same kind of developer philosophy, that users are dumb and they must be prevented from customizing the application has characterized the developers who have converted the Mozilla browser into Firefox, which is another unwelcome change that I have greatly hated.

[jorvi]

> The same kind of developer philosophy, that users are dumb and they must be prevented from customizing the application

The more holistic view is that every customization path incurs a support burden.

If you have 3 options to support, you can engineer and test the heck out of each option. 15 options? Not so much.

So having those 12 extra options not only creates permanent extra workload, as dev time is finite you’ve effectively made the 3 aforementioned options worse off.

[JohnFen]

> For me, indeed KDE 3.5 has been the best graphic desktop that I have ever seen.

I agree wholeheartedly with this. I miss KDE 3.5.

[account42]

Another problem with KDE 4 was the buzzword technologies pushed everywere - semantic desktop with Nepomuk (nice research project but not fit for normal use), plasma applet UI added to applications (why?), activities replacing virtual desktops but not really.

[bigstrat2003]

That version thing was frustrating because it was an unforced error. Surely someone, at some point, brought up that people would expect the version number to mean that it was ready for use. But they chose to proceed with using their idiosyncratic version scheme, and unsurprisingly suffered a reputation hit for it.

[FeepingCreature]

KDE4 ruined Amarok too. That's why there's Clementine and Strawberry now, which are forks from Amarok 1.4.

[cullmann]

Amarok did 'die' because the dev team dissolved due to multiple issues.

[FeepingCreature]

Huh, interesting. Well it still supports my point that when people looked to which version to resurrect, they did not pick Amarok 2.

[account42]

Amarok became unusable with KDE 4 before it died.