>You started with a *particular* change of direction in mind, one that
>was explicitly considered, and rejected, long long ago. Real-time OS
>design requires very careful ground-up work, and the Hurd servers do
>not have any interesting real-time properties.
From your narrations I'm not entirely sure whether you know what a real
time system is or what properties are beneficial for what I've discussed
(or even what I've discussed). I am not talking about LynxOS, or VxWorks,
or Chorus, and would never even consider that it is possible for the Hurd
can go in that direction; there doesn't seem to be expertise in signal
processing or real-time systems here.
What I talked about was a normal platform tuned to provide high throughput
and processing capabilities. I also said that, with interest, there would
be help (do you believe in teamwork?) Without interest, no one will care. I
tried to solicit opinion, foster a discussion and stir up a little interest
in the possibilities; apparently you don't believe in these things. That's
too bad. I was assuming that everyone working on this project would be open
minded to new ideas. Technical folks are a peculiar bunch after all ...
they tend to see things objectively.
Good luck to the Hurd. I've lost interest.