Re: about GNU Hurd

[Alfred M. Szmidt]

   How is it that you tell me that GNU Mach is not fundamentally
   obsolete, and then go on (in the next few e-mails you sent out) to
   say that Linux already does everything Mach+Hurd do, and more,
   without suffering from the same problems? It seems to me that is
   the definition of being obsolete.

You are confusing features, and the underlying design.  GNU Mach
mostly lacks drivers to be useful for most people.  It is software,
things can be added incrementally without having to reimplement
everything from scratch every odd year.

   However: Just because Linux is capable of surpassing a Hurd based
   system in its current (or even forseeable) capabilities, does not
   necessarily mean that Linux does this in the best way, or that it
   will continue to surpass potential future capabilities of the
   Hurd. Arguing against innovation because the status quo 'does
   everything we need' kills progress.

Ofcourse, nobody can predict into the future.  But that isn't what we
are doing, we are trying to finish the GNU system.  If that means that
Linux becomes obsolete in 20 years, so be it.  But the Hurd will
simply not catch up to the standards of Linux in these 20 years.  Be
it with Hurd-NG or with Hurd on Mach.

I should note that I simply do not care what kernel is picked, as long
as we stick to it; I just think that Linux is a realistic choice in
finishing the GNU system into a state that will have all the nice
things that we would like, the same applies to the Hurd on Mach though
that is alot more work.