[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Hurd? Re: Stopping KDE from becoming a standard



> I've been wondering lately... how hard would it be to make the
> HURD run on top of Linux (perhaps a modified Linux) rather than
> on top of Mach?
> 
> Do you think this might finally prove a solution to the "Lignux" problem?

Thus treating Linux as the "microkernel"...

The idea *appears* to have merit, although I can't claim competence to
realistically decide that.

It seems to be an interesting question: If the daemons that make up Hurd
were to be run atop Linux rather than Hurd (avoiding the issue that there'd
be some resulting code rewriting), what would be the major advantages and
disadvantage?

a) It has the benefit of perhaps allowing Hurd to actually be portable to a
lot of platforms.  (It *appears* that Hurd only runs on IA-32, although I
wouldn't want to push that point to strongly as if I say so there's bound to
be at least one person fiddling with a PPC or Alpha edition to contradict
me...)

b) It would certainly represent "friendliness" from the "people
perspective;" if Hurd ran atop Linux then there could be no argument that
"Those *?#&*(! Hurd people want us to stop using Linux!"

c) The resulting Hurd daemons might be portable enough that they could be
useful running atop other OSes.  It would be rather appalling if one could
run Hurd atop Windows NT, for instance :-).

I'd speculate that some of the following might prove problematic:
a) Linux may not be as efficient at IPC as Mach;
b) Linux may not be as efficient at memory management as Mach;
c) Confusion of roles between things running as "Hurd" processes as
compared to "native Linux" processes;
d) Linux changes too often (e.g. - the weekly kernel updates) compared 
to Mach;
e) It represents going in the wrong direction, moving from the ("we wish it
were small") Mach to Linux, rather than towards something that truly is
microscopic like L3/L4;
f) Obviously work would be involved in a move from Mach to Linux.

I can't put forcibly convincing arguments behind any of those; there may be
some other "killer" issue that I'm not thinking of.

--
Christopher B. Browne, cbbrowne@hex.net, chris_browne@sdt.com
Web: http://www.hex.net/~cbbrowne  SAP Basis Consultant, UNIX Guy
Windows NT - How to make a 100 MIPS Linux workstation perform like an 8 MHz 286


-- 
To UNSUBSCRIBE, email to debian-hurd-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org


Reply to: