Severin is right. The mule was overloaded.
I am happy to see that they all agree on the idea of “technical debt”.
Pressing a button and what is supposed to happen does not happen. Hiding button is just not a solution. If button was created, it was to be used. So, it is obvious that somebody’s workflow is impacted by that.
There is a temptation of denial of reality in developers world, sometimes. “If I can’t fix something, I will try to remove it. If nobody complains, today or during the week, that means that nobody uses it. One or two users are complaining, two weeks later. That is not so many people. They are conservative and don’t want to evolve. Then, months later, there are several dozen of users that are complaining.”
That attitude changes to “I put the thing on stand by status.”
I was not surprised that 2.80 was not delivered in 2018.
That is a coherent idea to refactor many big areas of Blender, at same moment, to install a coherence between them. I think that communication went well between developers. It is normal to confront developers and users point of views and to go through an iterative process to obtain that coherence.
Here, the denial was to proclaim that process will end-up after 6 months when it took 2 years for a stable 2.5 that did not contain so many tools.
I was surprised to how smooth feedback process went in last quarter of 2018.
At that moment, several wrong ideas were purged by users feedback. And, at same moment, technical debt relative to 2.8 designs started to accumulate itself. But 2.8 beta stayed mainly usable.
I did not really understand result of Homestretch in May. IMO, some UI changes pushed away 2.80 a little bit more, bugs were weirdly prioritized and importance of technical debt was probably underestimated, at that moment.
There are lots of bugs in bugtracker with a normal priority. When few conditions are just slightly different from default scene without being uncommon or rare conditions at all, a button, a feature just doesn’t work anymore.
Although, that makes 2.8 unreliable on many areas, it is incredible popular and many people are telling that they are using it exclusively. On many areas, 2.8 design is reducing technical debt of previous releases.
IMO, situation is going a lot better than during 2.5 as Jeroen said.
There will probably be some things that will force some users to keep 2.79 during a significant period. But if technical debt is significantly reduced in 2.81, only several months after 2.80 ; that will be impressive.
Clearly, they are talking about serious topics and long-term tasks that could not be solved in 2.81 and that could considered as technical debt older than 2.8.
So, I don’t expect 2.81 to solve every issues.
But I really expect that everything possible in 2.79, will be possible in 2.81.
That when user press a button, it simply does what it was introduced for.
I hope they can quantify technical debt, quickly.
I hope they can build a roadmap for 2.8 series that would satisfy users. If it means that release cycle should take more time, I think they should not hesitate to go to a 3,4 or 6 months cycle.
not afraid for a hang once in while, and i wouldnt mind if cycles would get a bit more complex to configure, all blender users becomme technical eventually so no need to over simplify. currently lots of attention goes to the GUI, maybe time to think about expansion GUI can still later be changed tuned (it wont stay forever like this, i guarantee, so why bother)