At 10:26 Uhr +0100 17.12.1997, wb2oyc wrote:
>>There is also Linux/mac68k (or some such name).  The last I heard it was
>>not ready for prime time.
>Yeah, but its worth checking from time to time.  They have been making
>good progress lately, except that right now their main web site is down
>for a major overhaul or something....

[...]

>The reason its interesting to me is this: shouldn't their port be
>able to provide better performance from a given machine, since they
>won't be utilizing any ROM'd code, or is that wrong?  If its wrong,
>why is that the case?

The usual story... for L*nux, all the world's an i386. The big difference
between the "Industry Standard Architecture" and the Macintosh is that the
Intel machines have the bare metal in common whereas the Macintosh unifies
a heterogenous hardware with a firmware layer.

Working around the ROMs as a principle is merely a religious issue, and
frankly, those who are the most excited about this idea usually rant and
then go elsewhere without having contributed anything.  I must have seen
about four or five longish threads on direct booting to BSD that never got
to any results; the efforts and stamina needed could be used elsewhere with
much greater effect.

"Performance" -- what performance do you expect to gain by "direct"
hardware access? We usually use the ROM code to find out about peculiar
hardware parameters or to let firmware code do jobs it knows best about,
e.g. for NuBus cards. If you don't have the specs for all the hardware out
there, how can you expect to do better than the card manufacturer or Apple
itself? Remember, this is not PeeCee crap - no Real vs. Protected vs. Blah
Mode issues here. And, seen from a CPU performance vs. OS performance
ratio, the Mac has never been too bad. I am pretty sure the low level stuff
in the ROMs is highly optimized.

IMHO, the MacL*nux folks still have to get real -- even more now that their
Mighty Warrior Alan Cox has left for the SGI port.

	hauke



--
"It's never straight up and down"     (DEVO)