>> Normal programs shouldn't be able to run in kernel space. And it should be >> able to recover from a badly written program. That's the purpose of memory >> protection that normal users are interested in. >> How did you mean that you can write a BSD program that craches the kernel >> without exploiting a kernel bug?
>drivers? almost all drivers in AmigaOS are 3rd party these days..... Perhaps I wasn't clear. I meant normal programs. Not drivers, or kernel code etc. With a good structured OS model with nice layers with good protection between them will minimize the amount of bugs introduced in the system that is fatal to more than just the program itself. The OS must consider that it's humans that program, and human tend to make more error the more complex the system is. My experence with windows has a worse model than for example Solaris based on the amount of time the total system colapses. Sure, there are less 3rd party drivers to screw up the system, but I still THINK that it's better done. I've seen how they handled message passing in Win 3.1 (for which Win is still backwards compitable to?! Or was that just Win '95?), and it was terrifying. Oh well. I haven't seen the inside of windows. But have seen many bad design choises coming from MS I'm quite sure that memory protection, resource tracking etc could be done in a much nicer way than the way MS does it. The craches can't all be blaimed on bad 3rd party programers. Hell, even cash machines with windows don't work like they should. All comment in the line of: "You don't know shit, so shut up" are not relevant and will be ignored. Counter examples and comments on the specifics itoh... / John
