[md8z]

I don't really know the details of this but for some things it is non-trivial to ship even small changes on other platforms, even those will have to go through full dev-test cycles which takes time and money. About removing things, I don't know, I had a car once that had some really nice cupholders. Really deep and exactly the right size, exactly at the right height for the arm. My current car doesn't have them and I can't find any cars that have that were quite as good as that. They just stopped making them. If that's not "vision" then what is it? I mean somebody has to make the decision of how to make the new cars. There is also the question of, do I value cupholders over everything else in the car? Would I buy a car with terrible steering if it meant the good cupholders? If I could get the good cupholders in any car would I pay an extra $3000 or however much the dealership charges? I think probably not but it really depends. So there's many factors at play here.

Edit: Or say maybe I am a startup founder and I design and build my own car exactly how I want and turn it into a company. It's perfect for me but then someday I get bored of driving my car and I retire. Then I hire someone else to design the cars and pass responsibility on to them and they change some stuff. Well, now the cars are different and everyone pretty much has to accept it because the original designer is gone, and as much as people liked the old one, nobody else can really copy them exactly because it was really their personal vision that made it what it was.

For Wayland, I think all of that is happening already? There is somebody making an implementation in Rust. They did try to make a shared implementation (Weston) but it turned out that people didn't actually want that, they preferred to write their own implementations.

[simion314]

The car examples don't make sense.

It is software, in Chrome case is just a simple popup, the code is there and it only is visible for OSX because the platform forced the vision guy's hand. The bullshit excuses that is hard to code and test and maintain do not work here.

Also excuses do not work even if valid if you destroy your users workflow, you don't remove system tray, server side decorations and just tell your users to find replacement applications because the ones they use do not conform to the GNOME vision.

Again, if is not a toy you target some users, is your duty to listen to this users and not to impose your vision on them, I am upset when there is no actual testing/research involving actual users and real world work, say when you test your app with "hello world" simple workflows that fail in real world with real users, or you make your app look cool on your expensive screen but looks like shit on real users hardware.

But you are right, GNOME has decided they don't want a part of the users and they are cultivating the perfect GNOME type user, a user that adapts to the software and not the reverse.

[md8z]

I don't know what you mean bullshit excuse. Everything has a cost to test and maintain. It doesn't help to say that it's bullshit if nobody has done a real cost analysis. Remember that this is something that has to be maintained for years. If some bugs occur in it later and it has to be removed again then the users will be upset again so it's not really useful for us to say just ship it and don't test, that's what we want to avoid. Yes, you and I could guess what it costs but that doesn't carry as much weight as somebody who actually works on it full time doing their cost analysis.

I get your frustration about your workflow but I'm still upset about my cupholders :) For the system tray and server side decorations, there are technical reasons for those to have gotten removed. Their existence may enable some workflows but it also breaks some other workflows so that's not an area where everyone can win. And if you want to bring them back then I can guarantee you that's not just a matter of flipping a switch, there is real work that needs to be done there and it won't happen if nobody is willing to pay the cost. It doesn't really make sense to blame volunteers for not being able to afford that either when this is something that's so expensive that the bigger contributors like Red Hat don't even want to pay for it.