Re: The idea of an own L4

[Alfred M\. Szmidt]

   However, it appears to me that the future of Hurd lies with
   l4-hurd.  Arhcitectural changes are happening in Hurd, and as this
   occurs the exisiting Hurd system will become progressively less
   important.

Some things will be changed in the Hurd to accomedate the Hurd/L4
port, but ditching all that has been written to this date won't
happen.  Alot of the code isn't tied to Mach at all.

As for where the future lies, I my crystal ball is misty...  Many
people are getting quite tired of the whole `lets wait for the FOO
microkernel before we continue porting'.  It is better to port now, to
whatever exists, and then port agains when the new versions comes to
existence.  This happened in the beginning of the Hurd project, TRIX
-> Mach -> TRIX or SPRITE -> Mach, and now it is happening again: ->
L4 -> L4.X2 -> L4sec or EROS or Coyotos or ...

This is more or less why I have completely ignored Hurd/L4, everyone
is afraid to take a decision and stick with it.  Just pick something,
and work on that...  Nothing is ever perfect.

   As an outsider, I do not hear people talking about "original hurd"
   much.  Perhaps I am listening in the wrong places. The successor
   has a good chance at greater impact.

The (original) Hurd is the only Hurd that people actually use, and is
quite acitvley maintained.  And frankly, I see a brighter future for
the Hurd on Mach than Hurd on L4 since there are more people doing
work on Mach than on L4 right now.

Cheers.