Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!seismo!sundc!pitstop!sun!pepper!cmcmanis
From: cmcmanis%pepper@Sun.COM (Chuck McManis)
Newsgroups: comp.os.minix,comp.sys.amiga
Subject: Re: Minix on Amiga ?
Message-ID: <32943@sun.uucp>
Date: Wed, 4-Nov-87 18:28:08 EST
Article-I.D.: sun.32943
Posted: Wed Nov  4 18:28:08 1987
Date-Received: Sat, 7-Nov-87 12:33:54 EST
References: <2156@gryphon.CTS.COM>
Sender: news@sun.uucp
Reply-To: cmcmanis@sun.UUCP (Chuck McManis)
Organization: Sun Microsystems, Mountain View
Lines: 37
Xref: mnetor comp.os.minix:1996 comp.sys.amiga:10443

In article <2156@gryphon.CTS.COM> jburnes@pnet02.cts.com (Jim Burnes) writes:
>Chuck:
>  I have been thinking about porting MINIX over more as a learning experience
>than as an attempt to improve on the Amiga Exec.  I am aware that the amiga Os
>is somewhat more sophiticated and also is designed to handle real-time events.
> Such is life.  If I port MINIX then just maybe I would understand a lot more
>about operating systems and more fully appreciate the amigas OS.  (OK..I'll
>admit it..I just think it would be plain fun!  ;-)  )

I guess I'm getting cynical in my old age :-). This is of course the one
reason why anyone would actually benefit from porting MINIX to the Amiga.
Yet, there are some tradeoffs and assumptions made in MINIX because they were
porting too a segmented architecture. I suggest reading the MINIX book,
Comer's book on XINU, the design of the UNIX operating system by Bach,
and P.B. Hansen's book on O/S design. Then with all those ideas brewing
in your head write your own O/S (which invariably falls into pieces like
Process control, I/O device abstraction, resource allocation, etc) and 
test it by replacing pieces of the Amiga OS. The beauty of the Amiga 
system is that it lets you do this in 'user mode' and all of the pieces
are accessible. In this way you can develop on your 'target' machine and 
still keep the benefits of an existing O/S around. Or use the pretested 
and working Amiga O/S to test out your new low level stuff. 

My personal experience is that while designing and writing a complete
O/S is too large a project to tackle, writing the pieces of one are 
managable. And after writing enough pieces you will suddenly find you
have written the whole thing. The benefit of reading all those books
is that you will know some of the problems that the O/S has to deal
with and that can help you understand why certain parts of the system
do things in what might otherwise be inexplicable ways. 

Now after you have done your homework send me your resume :-)


--Chuck McManis
uucp: {anywhere}!sun!cmcmanis   BIX: cmcmanis  ARPAnet: cmcmanis@sun.com
These opinions are my own and no one elses, but you knew that didn't you.