Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!husc6!hao!noao!arizona!rogerh
From: rogerh@arizona.edu (Roger Hayes)
Newsgroups: comp.sys.amiga
Subject: The Next Generation
Message-ID: <2785@megaron.arizona.edu>
Date: Thu, 12-Nov-87 14:05:20 EST
Article-I.D.: megaron.2785
Posted: Thu Nov 12 14:05:20 1987
Date-Received: Sat, 14-Nov-87 19:08:18 EST
Organization: U of Arizona CS Dept, Tucson
Lines: 32
Summary: no virtual memory!

There have been some folks clamoring for virtual memory for the next
generation.  I would like to speak up on the other side of the issue.
Some of the things that makes the Amiga so nice to use and to program for
are the extremely low latency of task switching and the extremely low
overhead of message passing.  Both of these are likely to go down the
drain if memory protection is added.

As an extreme example, look at Sun workstations -- it takes tens of
seconds to open a window, because you have to make several trips down
through Unix and back up into the application code to get the job
done.  On my Sun at work (a 2-50), I can easily drag a window faster
than the display can keep up, so that the window outline chases around
the screen for 10 seconds after the mouse is still.  This is because
Sun task switching is slow -- thanks to Unix and virtual memory.

The latest work in distributed operating systems is tending towards a
lightweight process model -- just what we have in the Amiga.  This is,
IMHO, because a distributed OS is at heart a real-time I/O handler for
the network and display devices.   The Amiga does a great job at this
now.  Let's not lose it in an effort to gain ground that the pioneers
are deserting.

You want more security?  Well, I might accept memory *protection*
hardware, as long as the additional cost in task-switching was
recoverable from elsewhere.  Or we could make the software do it --
write programs in a secure language.  This is similar to the RISC
approach to pipeline delays.  The rationale is that compilation is a
one-time cost, while hardware complexity hurts every time you run a
program.  (And that's disregarding the production cost entirely.)

        Roger Hayes
        rogerh@arizona.edu