Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!cs.utexas.edu!ted
From: ted@cs.utexas.edu (Ted Woodward)
Newsgroups: comp.sys.mac.system
Subject: Re: System 8 (What enhancements do people want?)
Message-ID: <701@needmore.cs.utexas.edu>
Date: 5 May 90 07:13:21 GMT
References: <1990Apr20.193426.16869@uswmrg2.UUCP> <5566@okstate.UUCP> <1990Apr25 <67@victoria.cs.utexas.edu> <2129@zipeecs.umich.edu>
Organization: U. Texas CS Dept., Austin, Texas
Lines: 24

In article <2129@zipeecs.umich.edu> gilgalad@dip.eecs.umich.edu (Ralph Seguin) writes:
 
>Suppose you want to do a large ray trace?  I wouldn't want to sit around for
>several hours without being able to use my system for something else.
>With preemption, you just shove the process in the background at a low
>priority, and it gobbles up any CPU time that you do not use.  Generally, when
>you are doing something like word processing, there are tons of cycles to spare
>since you appear to be moving about as fast as granite to a computer.

If I were writing a ray tracing program, I would call WaitNextEvent() every
so often (note C bias :-) ).  This allows the background raytracer to give
control back to the foreground application.  Then you could use the system for
whatever else you wanted.  Designed correctly, applications like this work
very well in a cooperative multitasking environment.  You just have to be
willing to give control back every so often; often enough that the user
has a minimum of frustration because the Mac does not respond (ie, NOT
Printmonitor...:-( )



-- 
Ted Woodward (ted@cs.utexas.edu)

Someone shot the food...