Path: utzoo!mnetor!uunet!husc6!hao!gatech!mcnc!decvax!ucbvax!CORY.BERKELEY.EDU!dillon
From: dillon@CORY.BERKELEY.EDU (Matt Dillon)
Newsgroups: comp.sys.atari.st
Subject: Re: Multitasking - A nightmare
Message-ID: <8712312200.AA15560@cory.Berkeley.EDU>
Date: 31 Dec 87 22:00:31 GMT
Sender: daemon@ucbvax.BERKELEY.EDU
Lines: 56

:  I think that this discussion is a little off-base in many ways. Would 
:this be going on if we were debating the benefits of disk versus cassette 
:storage?
:  You will be hard pressed to find ANY microcomputer of recent vintage 
:that does NOT multitask to some extent. The Atari 8bits use multitasking.
:The ANTIC CPU interrupts the 6502 chip to generate the video. Smart programs
:turn OFF the multitasking during cpu intensive stuff (compiles for example)
:to speed things up, or write special multitasking programs (graphic 
:displays). The POKEY sound chip works in a similar manner for the sound 
:output. Disk controllers use dedicated chips (and DMA) to do their 

	Ever heard of newlines?  You paragraphs were all one single line!
I think *you* are slightly off base a little here.  Every microcomputer down h
to the old Apple II's (and older) has interrupts.  Interrupts and DMA does not 
a multitasking computer make.  I think the definition everybody has had in mind
throughout this affair has been:

	-preemptive scheduling
	-each 'task' is an independant entity with no restrictions relative to
	 other tasks (Read: Not an interrupt, Not DMA)
	-The OS supports the fact with signaling, control, and communications
	 calls, is reentrant, etc...

	The overall idea being that you can then run completely independant
applications simultaniously.  After all, any semi intelligent hacker can write
a simple context switcher for his or her particular program.

	Now this business of running more than one application at a time
has been argued to be not all that big a deal.  Even discarding the rather
silly arguments over exactly what one would want to run simultaniously,
multitasking has distinct advantages in other areas.  Take a look at the
number of problems multitasking simplifies (things you CAN DO NOW with hacks or
complex programming):

	-DA's (graphical clocks, calculators, etc...)
	-Printer spoolers
	-Networking
	-Asynchronous IO
	-Multi-line BBS's **without** delays
	-program debugging
	-OS hacks (hot keys and other items)

	Looking at the amount of money and time Apple has invested 
attempting to match the Amiga's multitasking.  Even if you don't give a
bull's toe for the Amiga it should be obvious that the future of 
microcomputers is turning towards multitasking.  Look at (Bletch) OS II for
the IBM!  If anybody has ever had a rock solid will-never-change attitude it
has been IBM.

	As I said, the above can be done NOW on the Atari, but it takes a lot
of work, and many incompatibilities crop due to the lack of a standard way of
doing some of the above.  I think that simply adding a couple OS calls to
support a rudimentry multitasking (to go beyond Apple's task switching) would
be a big boon!  The first step is to share the CPU.

					-Matt