On Mon, 14 Mar 1994, Nicholas Mitchell wrote:

> Is the professional version as "functional" as A/UX would be
> (assuming you get "MachTen Prof." plus "MachTen X windows package"

Not even close.  MachTen is like a unix version of "SoftPC".

It is an emulator, not an operating system.  It runs as a 
Macintosh application on top of System 7, which lacks memory
protection and preemptive multitasking.  This means that 
a Mac running MachTen is as vulnerable to crashing because
of a buggy INIT or Mac app as any other Macintosh.

It's principle virtue stems from its chief flaw.  Because
it runs on top of MacOS, and shares its filesystems, it 
will run on many Macs (especially PowerBooks) that aren't 
supported by A/UX, with less dedicated disk, and less CPU.
There is even a version that runs on machines without an MMU.

MachTen is undoubtedly a well-polished package, but it is 
going NOWHERE, as far as I can see.  People with an interest
in a MachTen-like working environment who are also interested
in contributing to the future of free software would do better
to support the on-going development of MacMiNT, which is a
free Mac port of a unix emulator originally written for 
Atari TOS.

> (i.e. A/uX seems to be at the boot of apple, 
>  MacBSD doesn't seem like it'll support my centris 650 for quite a while
>  and Tenon seems to have some sort of future)

Maybe so.  But if you want to _use_ unix on the Mac, as opposed
to kernel hacking toward the eventual goal of a freely
redistributable unix on the Mac, A/UX will be the best choice
for a C650 for the reasonably near future.  It's also a 
reasonable box for doing MacBSD cross-development, I'd 
imagine, so buying it doesn't lock you out of making a
contribution.

What would be a waste of time, in my view, is to invest
money and effort in learning, maintaining, administering,
and porting software to MachTen -- a proprietary unix
emulator that is used by a relatively small number of 
people.

I've appended a similar comment that I made on the MacMiNT 
mailing list recently.  "alt.mac.os@tazboy.Jpl.Nasa.Gov"
is the name of that list, for those who want to learn 
more.  Fairly recent ports of MacMiNT are available on
sumex and umich, I believe.  The readme files in those
archives will point interested folks toward more info
and software that is available for MacMiNT.

- Bill

/* Please note that I'm not trying to slam Tenon or Anita
   for making a special offer to those interested in MacBSD.
   I admire clever software, regardless of whether it's 
   free or proprietary.  It's good business for Tenon
   to target MacBSD folks;  after all, I paid for A/UX,
   and I still follow MacBSD and MacMiNT developments.
   However, my bias is toward advancing the state of the art 
   for the long run, an objective that I think is better 
   served by the minority now hacking away on free software 
   with freely available source.  */

From johnston@me.udel.edu Tue Mar 15 00:02:31 1994
Date: Fri, 11 Mar 1994 15:01:33 -0500 (EST)
From: Bill Johnston <johnston@me.udel.edu>
To: Jonathan Birge <jbirge1@cc.swarthmore.edu>
Cc: alt.mac.os@tazboy.Jpl.Nasa.Gov
Subject: Re: MacMiNT Installer

> Also, is anyone running MachTen? I'm thinking of getting the development
> package for it, but it just occurred to me that since gcc is free, if
> someone compiled gcc under MachTen they could make it publicly available.

This does not make the MachTen c libraries free, so the fact that you
can copy the compiler is of limited value.  What good is having a 
compiler if you can't link them to create executables?

This wasn't traditionally an issue for unix machines, because no one
really considered that machines would be sold with compilers and 
libraries unbundled. 

What you could do is to try to get someone to port glibc to Machten,
and freely distribute that along with gcc.  The difficulty there
would be in obtaining enough low-level info about the way unix
system calls are made under Machten, and in doing so in such a
way (ie, without having to sign a non-disclosure agreement with
Tenon) that one could arrive at an unencombered port of glibc.

Frankly, I wouldn't waste alot of time porting code to a 
proprietary unix emulator that is used by a fairly small
number of people.  If you want your work to carry on to 
future platforms, I think it makes more sense to support
MacMINT and MacBSD, or more sensibly (IMO) NetBSD or Linux
for the intel x86 architecture.

After a few years of fighting just to keep running A/UX on 
my Mac with a decent variety of net.software, I've come
to the conclusion that it's more sensible to use the Mac
as a front end, connecting to external unix boxes running
on better-supported hardware.  

More power to the guys working on MacBSD and MacMINT, 
but in the long run I think that the state of the art 
will be better served by focusing effort on open architectures. 
More power to Apple if it can maintain market share by continued 
innovation and idiot-proof-ness, but why fight them for hardware 
info that they don't seem to want to share?

- Bill







------------------------------------------------------------------------------