CSD vs SSD
I have been reading and thinking about CSD vs SSD a long time and this is what I think:
CSD
+Flexibility
+Simplifies Wayland implementations
+Cleaner speration of conserns, a compositor shouldn't do any rendering of gui
-Can't impose graphical conformity regarding to decorations, the user experience can be confusing if every Application/toolkit chooses to go their own way
+Better model for a voluntary graphical conformity , a system "GUI theming and settings" system that is toolkit agnostic giving applications and toolkits information/hints how to render in a conforming manner
SSD
+Impose Conformity and make the user experience less confused and integrated
-Adds complexity for the Wayland implementation and do break separtion of conserns
-Harder to implement custom GUI design - for a minority applications and use cases the "GUI defaults" are not an optimal solution
I think KDE's decision to use SSD in KDE 5 may be the right and rational decision for them (and only for them). I don't know much about KDE but all software projects live with a legacy, as I understand KDE has always tried to achive a nice conformity in the user experience, even with "alien" toolkits like gtk+. For KDE to rewrite everything and change core principles in a short time span and such new and unknown technology as Wayland, would be impossible. Maybe in KDE 6 there are a time for such work.
I have been reading and thinking about CSD vs SSD a long time and this is what I think:
CSD
+Flexibility
+Simplifies Wayland implementations
+Cleaner speration of conserns, a compositor shouldn't do any rendering of gui
-Can't impose graphical conformity regarding to decorations, the user experience can be confusing if every Application/toolkit chooses to go their own way
+Better model for a voluntary graphical conformity , a system "GUI theming and settings" system that is toolkit agnostic giving applications and toolkits information/hints how to render in a conforming manner
SSD
+Impose Conformity and make the user experience less confused and integrated
-Adds complexity for the Wayland implementation and do break separtion of conserns
-Harder to implement custom GUI design - for a minority applications and use cases the "GUI defaults" are not an optimal solution
I think KDE's decision to use SSD in KDE 5 may be the right and rational decision for them (and only for them). I don't know much about KDE but all software projects live with a legacy, as I understand KDE has always tried to achive a nice conformity in the user experience, even with "alien" toolkits like gtk+. For KDE to rewrite everything and change core principles in a short time span and such new and unknown technology as Wayland, would be impossible. Maybe in KDE 6 there are a time for such work.
Comment