Re: Supporting POSIX *users*

[Jonathan S. Shapiro]

On Tue, 2005-11-01 at 17:49 +0100, Alfred M. Szmidt wrote:
>    Yes, but the *really* cool thing about a secure system is that even
>    when the user does this, they can't screw *me* up.
> 
> That is how a POSIX system works, so it isn't true only for `secure
> systems' (I'm assuming that a system that follows POSIX like GNU isn't
> a `secure system').

This certainly has not been my experience. My experience has been that
there is only one type of POSIX compromise in practice: total
compromise.