NetNews Usenet Archive 1992 #18/NN_1992_18.iso/spool/comp/editors/2023

Internet Message Format | 1992-08-18 | 3.9 KB

Xref: sparky comp.editors:2023 comp.os.minix:4233 comp.os.linux:8539
Path: sparky!uunet!mcsun!sun4nl!star.cs.vu.nl!philip
From: [e-mail] (Philip Homburg)
Newsgroups: comp.editors,comp.os.minix,comp.os.linux
Subject: Re: New ftp location for the folding editor Origami
Keywords: origami
Message-ID: <[e-mail]>
Date: 18 Aug 92 13:53:30 GMT
References: <[e-mail]> <[e-mail]> <[e-mail]>
Sender: [e-mail]
Followup-To: comp.os.minix
Organization: Fac. Wiskunde & Informatica, Vrije Universiteit, Amsterdam
Lines: 70

In article <[e-mail]> [e-mail] (Henry G. Juengst) writes:
%In article <[e-mail]>, [e-mail] (Philip Homburg) writes:
%>
%> I think that if you can keep Origami more or less POSIX compliant, porting
%> it to a (my) 32 bit version of Minix is not much work. At least the diffs
%         ~~~~~~
%Very nice. What do you think how long will anybody wait until the official
%MINIX does have features like virtual memory, long file names, network
%file system, X11 etc. ? I can't even use these technics today under MINIX
%if I would buy a much better system (not 8088 PC). And why should I pay for
%a new compiler, libraries and bug (!) patches again and again ?

FLAME ON

First of all, this is my opinion and necessarily someone else's.
I think that Minix is for people who want to study operating systems
and for kernel hackers (Unfortunately, if you a mc68000 based system, there
are not many other Unix like systems besides Minix). If you have an
application you want to run, you either patch the application to use the
features your OS offers, or you change your OS according to the application.
There is no requirement of running applications out of the box.

I think that Andy decided to keep Minix simple because students either
don't have the hardware to run virtual memory (PCs, ATs, Amigas, Ataris, Macs),
a network file system, or X, and that they probably don't understand the
system with these features (did you ever study the way paging interacts with
NFS for example).

This is also the reason I play with Minix, I wanted to implement some of these
basic OS concepts like VM, TCP/IP, do an X port, etc.

Of course there is no reason why you should by a new compiler. You can write
your own compiler (like Bruce Evans did), buy enough hardware to run gcc,
or get access to a proprietary compiler (like I did).
Bug fixes are a fact of life, but people even complain when they only have to
get a new set of Linux binaries from an ftp site. And of course you can always
ignore bug fixes.

%POSIX ? I've got MINIX 1.5.10 (plus patches up to 1.6.16) for a PC, but if
%it has to eat some POSIX source it says after some hours "!@#$%^&*()+". I
%will remove MINIX from its hard disk some days later and just use it to play
%chess and as silly terminal. A much better system (hardware + OS) is on
%its way. Please, don't tell me anything about any patches. I know them all
%and can't see them any more. MINIX _was_ a nice idea. Sorry.

I think that I can't develop new Minix versions if I only had a PC, I'm just
to lazy. I'm used to get core dumps when a process dereferences a NULL pointer,
compiling large ANSI-C sources, etc. That part of Minix (running on small
systems) was a nice idea, and it worked for a number of years (I remember me
recompiling Minix on an XT with one floppy drive, on a Minix kernel with a AT
harddisk driver). But since you want VM, X, large filenames, you have to have
a much larger system than a PC or an AT. And then it is only matter of
choice: no kernel hacking take SVR4, maybe Linux or 386BSD, lots of kernel
hacking take Minix.

FLAME OFF

I redirected follow ups to comp.os.minix




                        Philip



Philip Homburg                            <[e-mail]>
Vrije Universiteit / Dept. of Math. & Comp. Sci.    [phone]
Amoeba project / De Boelelaan 1081A
1081 HV Amsterdam / The Netherlands