FXDDR@ALASKA.BITNET (02/10/88)
This isn't strictly related to the ST, but I thought that the people who
recently argued that civilization, yea verily life itself, could not
continue without multitasking micros might find this amusing. It's from
the "Program Notes" column of the Feb 88 IEEE Spectrum (p19).
Not-quite-MacMultitasking
...
An August press release from Apple Computer Inc. trumpeted the release
of MultiFinder, the "first multitasking operating system for the
Macintosh." But the software might be more appropriately called
cooperative system-sharing.
... (several paragraphs about the restrictions and catches in
MultiFinder) ...
Even "Program Notes" was taken in; the column [Spectrum, December 1987,
p. 22] claimed MultiFinder was an actual multitasking system. Apple
says, however, that operating systems should be driven by the
applications' needs, so perhaps true multitasking for the Mac will be
relegated to A/UX -- Apple's version of Unix.
From the list of limitations they quote, it sounds like Apple has a ways to
go before it catches up to Beckmeyer's line...
I still don't like the wimpy pseudo-multitasking like Amiga and Unix push.
I prefer real multitasking, where each process has its own processor!
But I've got a lot more soldering before I get there...
Don Rice
FXDDR@ALASKA.bitnetharald@leo.UUCP ( Harald Milne) (02/12/88)
In article <8802100646.AA16770@ucbvax.Berkeley.EDU>, FXDDR@ALASKA.BITNET writes: > I still don't like the wimpy pseudo-multitasking like Amiga and Unix push. > I prefer real multitasking, where each process has its own processor! > But I've got a lot more soldering before I get there... Well at last count, my Amiga has 3 CPU's and 6 coprocessors. Yes, they run simultainiously. Oh wait, you mean multi-processing. Well, when UNIX does this, with three disparete environments, let me know! No problem multi-tasking. Just have to fight for 2 Ethernet addresses for bandwidth. -- Work: Computer Consoles Inc. (CCI), Advanced Development Group (ADG) Irvine, CA (RISCy business! Home of Regulus and hamiga) UUCP: uunet!ccicpg!leo!harald
pes@ux63.bath.ac.uk (Smee) (02/16/88)
Oh golly, I wish you hadn't said that. Still, long as you mention it, I don't mind multi-processes sharing processors. What I object to is them sharing memory and peripherals. When someone comes up with a 'home' micro with individual protected memory 'partitions' for each process, with all inter-process communications and all external device use forced through (secure gated) entries in the O/S, *then* I will believe in multi-tasking on home machines...
ross@swan.ulowell.edu (Ross Miller) (02/17/88)
In article <1894@leo.UUCP> harald@leo.UUCP ( Harald Milne) writes: >In article <8802100646.AA16770@ucbvax.Berkeley.EDU>, FXDDR@ALASKA.BITNET writes: >> I still don't like the wimpy pseudo-multitasking like Amiga and Unix push. >> I prefer real multitasking, where each process has its own processor! >> But I've got a lot more soldering before I get there... > > Well at last count, my Amiga has 3 CPU's and 6 coprocessors. Yes, they >run simultainiously. > > Oh wait, you mean multi-processing. Well, when UNIX does this, with >three disparete environments, let me know! Hmmm, well Sequent does run multiple processes on multiple processors and unix works fine. Alliant doesn't seem to have a problem mixing unix up with CE's and IOP's. I believe the IOPs are 68020's. CDC 825 (don't know if this environment is still for sale) ran NOS, NOS/VE, and an unix sys V clone all at once. Note that the processor was switching instruction sets and memory managment schemes in mid stream. Quite a nice idea, and it worked. The hardware was great, but the software was a different matter entirely. I remember one of the engineers looking at incredously when I asked for wild cards in the SCL for NOS/VE. The Amy is nice technology for personal computers, but it is by no means the forfront of technology today. Ross -- csnet: ross@ulowell.cs.ulowell.edu uucp: ross@ulowell.cs.ulowell.edu || ...harvard!ulowell!ross Trust the computer. The computer is your friend.
jejones@mcrware.UUCP (James Jones) (02/20/88)
In article <2170@bath63.ux63.bath.ac.uk>, pes@ux63.bath.ac.uk (Smee) writes: > When someone comes up with a 'home' > micro with individual protected memory 'partitions' for each process, with > all inter-process communications and all external device use forced > through (secure gated) entries in the O/S, *then* I will believe in > multi-tasking on home machines... You mean like one of the GIMIX 6809-based SS-50C bus computers? The price may be a bit steep for "home" micros, but they've been around for years, and the memory management does have teeth in it. OS-9/6809 Level Two runs quite nicely on them. I don't know enough about the GIME chip in the CoCo 3 to be able to say how protected processes are from one another--they *do* have separate address spaces, though, and the price is certainly in the "home" computer range. You might want to pay a visit to someone who owns one and runs OS-9 on it. James Jones [This message has been treated with new, lemon-freshened DISCLAIMO, which renders it free of any opinions of any organization--save the multicellular one that posted the message, of course.]