I think a general problem in KDE (may be true for other software projects, too) is that the developers are always driven by stuff that's challenging and fun to implement, but they (understandably) don't enjoy taking care of the "easier" but more time consuming details. They start out with great plans and change everything from scratch, but they totally underestimate how much work it will actually be to get everything right again, so they run out of time to do it.
The loser in all of this is the end user, who wants to have a working desktop. Whenever it is working "too well", the developers come up with some great new plan how to do it all better. But in reality, this means that 80%-90% of the time the users will have to put up with a system that is broken to varying degrees. Seems just like a broken design model to begin with. It was like this in both KDE3=>4 and 4=>5 transitions.
And when the users complain, they are told that it's all the work of volunteers and that they should not have any expectations. They are told to fix it themselves, because it's all FOSS. But from a user's point of view, one could say that it's the developers who actually broke it. The ambitious developers get to steer a project into a certain direction, even if the community suffers from it. That's a very fundamental unsolved problem.