I'd rather have a more sparse, bare bones OS like OS4 and have the option of adding programs to do things rather than have it built into the OS.
Virtually everything you suggested there Kamelito, I'd never use, it'd be more bloat added to OS4. There's add on programs that do most of that stuff if you wanted it, and all without bloating out a trim, lean little OS.
Now that you mention lean, I'd like to point out that the current Workbench is about as far away from that goal as you can possibly get whilst at the same time giving the appearance of offering a restricted set of functionality. It may seem simple and bare-bones, but not by design, only by accident.
Workbench today is built upon code which reaches back to 1985, when it was created under great pressure. That code was overhauled in 1989/1990, reworked in 1998/1999 and has seen numerous changes over the following decade. What did not change in all this time is the basic architecture, which very much restricts what changes are possible in the first place. At its heart the Workbench today is still very much the same thing at it was 30 years ago. What Workbench does today ought to be done with less effort and resources spent: the results should be reflected in how they came about.
I tried my best back in 1998/1999 to open up the Workbench APIs, rework the input and refresh logic, allow for operations to run in parallel, but there was only so much you could do. Workbench is like the world's biggest ball of twine, except when you pull hard on the twine, the ball doesn't unravel, it actually pulls you back with twice the force.
