Path: utzoo!utgpu!news-server.csri.toronto.edu!bonnie.concordia.ca!thunder.mcrcim.mcgill.edu!snorkelwacker.mit.edu!apple!heksterb
From: heksterb@Apple.COM (Ben Hekster)
Newsgroups: comp.sys.mac.system
Subject: Re: True Multitasking
Summary: System resources should not be preempted by the OS
Message-ID: <48134@apple.Apple.COM>
Date: 16 Jan 91 00:16:46 GMT
References: <19019@shlump.nac.dec.com> <42588@ut-emx.uucp> <1991Jan15.233452.1163@maverick.ksu.ksu.edu>
Organization: Apple Computer Inc., Cupertino, CA
Lines: 43

awessels@ccwf.cc.utexas.edu (Allen Wessels) writes:

>Every company that sells an OS has programming guidelines, and I'd guess that
>most of them will tell you the bets are off if the program is written in an
>OS-unfriendly fashion.

jxf@orion.cis.ksu.edu (Jerry Frain) responds:

> OS-unfriendly?  Now there's a nifty term.  A program should have to be
> "OS-friendly," that's a lot of what this whole discussion is about.
> 
> A primary job of the typical operating system is to  manage resources.
> Memory, disks, processes, etc. are all resources.  If an OS cannot manage
> a resource effectively and efficiently, then that part of the OS should
> be enhanced so that it can manage the resource properly.

These are indeed all examples of system resources.  Although the OS does manage
resources, it is up to the application to use them wisely.

	An analog to the processor-preemptive multitasking system, we could
have memory-preemptive multitasking too--use all the memory you want, but
keep it too long, and the OS pulls the rug out from under your feet...  Save
documents on disk?  Fine--but if some other application needs space on a full
volume, the OS will automatically delete enough files to make room for it...
Well, you get my drift.  To me, processor non-preemption seems to make sense.

	In the same way, still my opinion of course, if an application
programmer feels that the user will be best served by completing a certain
task as quickly as possible without interference from other applications that
may happen to be present, he can implement this.  If the application is
performing a lengthy task, it might relinquish processor time to allow its
user to do something else in the meantime.

	Of course this means that bad programs (that do not call WaitNextEvent
as often as they might) will hog processor resources.  So will applications
that make extravagant NewHandle calls hog memory resources (and some do!).

_______________________________________________________________________________
Ben Hekster                           | "I don't want to start any blasphemous
Student intern                        |  rumors/ but I think that God's got a
AppleLink:   heksterb                 |  sick sense of humor/ and when I die
Internet:    heksterb@apple.com       |  I expect to find Him laughing"
BITNET:      heksterb@henut5          | --Depeche Mode [Some Great Reward]