> Neither one [kernfs & procfs] is truly necessary for system
> operation, although they DO appear to be performance treats, and
> they do allow for perusal via ls/stat/cat as opposed to needing
> special permissions thru the kernel (standard owner/group
> permissions seem to apply).

One advantage I like is that executables that use procfs instead of
groveling around kernel structs tend to continue to work even though
kernel internals have been changed.  The same is true for debuggers
that use /procfs (although I understand our procfs isn't up to that
task).

As much as I'd like to make procfs required, I can't justify endorsing
that proposal until we have utilities that can make use of it.  

	--jtc