Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!cs.utexas.edu!uunet!munnari.oz.au!cs.mu.oz.au!ok
From: ok@cs.mu.oz.au (Richard O'Keefe)
Newsgroups: comp.os.minix
Subject: Re: dosread.c again
Message-ID: <2492@munnari.oz.au>
Date: 23 Oct 89 03:09:06 GMT
References: <695.254152F7@mudos.ann-arbor.mi.us> <3768@ast.cs.vu.nl>
Sender: news@cs.mu.oz.au
Lines: 25

In article <3768@ast.cs.vu.nl>, ast@cs.vu.nl (Andy Tanenbaum) writes:
>  2. Virtual memory.  I tend to regard this as obsolete.  With Bruce Evans'
>     protected mode kernel and a 2M 386 you can have up to 2M of programs
>     running at once.  That has to be enough for a personal computer.

Um.  I'm using a machine with 4M.  The manufacturer doesn't make them that
small any more.  Let's just stop and calculate for a second.  How big a
square matrix of IEEE doubles will fit into 2M?  N**2 * 8 = 2 * 1024**2
means N = 512.  (The "2M of programs running at once" has to include the
data they are currently working with!)  Ok, linear programming on a
personal computer is out.  How about Lisp?  Well, 'size $OAKLISP' reckons
the total is 400k, and the bootfile it loads is another 400k.  So that
lets me run two programs the size of OakLisp on a personal computer.
Not a lot.  GNU Emacs runs fine on 386/UNIX systems with 4M of memory,
but you don't really stand a chance on 2M without VM.  Heck, even Jove
starts at around 300k.  How big is a statistics package like GENSTAT or
S?  Is there any reason why I _shouldn't_ want to run them on a "personal
computer"?  "2M has to be enough for a personal computer"?
	
I'm not knocking Minix.  Nobody ever claimed that Minix was the answer to
everyone's needs.  There are other more urgent things to do to Minix.
(Like upgrading the ST version.  It just so happens that I have an ST at
home...)  There are good reasons for not doing VM for a while (like not
putting more 80*86 dependencies in and leaving the ST further behind...).
But "2M has to be enough for a personal computer" is not a good reason.