Yep. :-). I've been thinking about this ever since JMF and TW did SMP. No disruption or no restarting? The former may be impossible...
Reload? Or do you mean replace? A reload loses all the ongoing data such as current program counters of each process, contexts, minute but important counters as date-time stamps.
They should first work on suspend as the next step for this.
Now will you go look at any PDP-10 OS? I can't imagine any VMS having this restriction. I can't remember not being able to replacing EXEs on the disk while people were running them.
This such a bug I can't begin to tell you how bad it is.
No, it doesn't. Take a look at a real operating system. Like I said, MS-DOS is an example of exactly what not to do.
Yea, it's degenerated to the lowest of the low. I wonder if this is a side effect of having the User Device From Hell as the common user interface.
Not if the trouble happens every time you look at the system.
Troubleshooting requires a modified form of the Scientific Method. First you verify you can reproduce it. Then you start to remove everything that doesn't affect the problem. Then you start to troubleshoot. buttuming you can identify the problem, then you have to figure out if you are going to fix it, bandage it, declare it an buttpect, or declare it a feature. In many cases, you won't figure out the real cause and produce the real fix until many, many, many years later.