[KLUG Advocacy] Re: BSD Amiga vs BSD Mac

Mike Williams advocacy@kalamazoolinux.org
Mon, 19 Aug 2002 13:47:15 -0400


On Mon, 19 Aug 2002 12:00:01 -0400
> 
> >If it
> >was fairly generic 680x0 code, wouldn't it run on a Mac IIvx at almost
> >the same speed as on an Amiga?
> 
> Don't forget that 68xxx series Macs don't even have hardware primitives
> for putting text on the screen (like the PC Video BIOS does, and EVERY
> SINGLE VIDEO CARD EVER MADE!).  It is ***ALL*** done in software on the
> 'early' Mac 68xxx boxes.  Scary.

Yuck!!  I know things like the rounded rectangle buttons were in ROM,
but what kind of idiot designs a graphical interface (from the start,
not a retrofit like MS Windows) makes the processor handle ALL the
graphics chores?
 
> >>(Nothing against Apple,  or the utility of these machines as what they
> >>were.  But a spade is a spade;  those machines were crap.  I actually
> >>read through the hardware specs once...... ick......  Useful as a
> >>workstation running a preemptive multi-user POSIX-ish OS?  No way!)
> > Most of the Amiga's fancy chips were devoted to multimedia and I/O if I
> >remember correctly.  I've never read through the specs of either machine
> >in any detail, but with the exception of Agnes handling some memory
> >access chores, 
> 
> Don't discount "some memory access chores".  How an architecture handles
> memory access and moving block of bytes into and out of devices matters
> a heck of a lot more than CPU MHz.

I had a feeling that was gonna make most of the difference.  Agnes, even
in her earliest iteration, was a pretty smart chip.

> Lets take my 90MHz IBM RS/6000 which purred away under the load of ~120
> users, and replace that with a 90MHz Pentium and the same amount of
> RAM.  We'll use the same smart SCSI controllers.... Ha! Yeah right.

Don't know much about IBM machines, but from their cost and stability 
(2 years between reboots and such!) I imagine they're pretty smart
inside.

> >how much would a text-based server use them?  I guess
> 
> Think "no graphic primitives in hardware" Ahhhhhhh!  Press the letter
> a... and the CPU says "OK! now to draw the letter A we must...."

... put a dot here, and another one next to it...Ack!! that's awful.  Even a "glorified word processor" should do better than that.
 
> >there's hard drive I/O, but the IDE-based Amigas (4000, mostly) had
> >pretty brainless IDE controllers.  
> 
> "Intelligent" IDE is a very recent event.  This was clearly a
> step-backward-to-try-and-be-cheap death gasp on the part of Commodore.

Yeah, it was a desperation move.  Quite a few of the hard drive
maintenance tools in the OS actually called them SCSI, so it's fairly
clear they were supposed to be.  And every other hard drive controller
for the Amiga platform drove SCSI.

I just remembered that Commodore made a few Amiga 3000UX machines that
actually ran real Unix.  This was long before Torvalds proved that Unix
was possible on less than a mainframe.  I think these machines were
stock 3000's with a network card and some extra memory, but I've never
actually seen one.  Anybody ever touched one of these freaks of
engineering?

-- 
Mike Williams <knightperson@zuzax.com>