Today, Andy Tanenbaum has officially announced the release of MINIX 3.0, the third stable version of this rather legendary operating system. The launch of v3 has been accompanied by a new website and a new logo. From the new website: “MINIX 3 is a new open-source operating system designed to be highly reliable and secure. It is based somewhat on previous versions of MINIX, but is fundamentally different in many key ways. MINIX 1 and 2 were intended as teaching tools; MINIX 3 adds the new goal of being usable as a serious system on resource-limited and embedded computers and for applications requiring high reliability.” Read on for more information.
In this announcement, posted in comp.os.minix June this year, Andy Tanenbaum announced he and his group were working on an updated version of MINIX, which had its last major release in 1996 with version 2.0. However, in an email conversation, Andy Tanenbaum asked me not to announce this; he did not want the press all over it until the official release, planned for the end of October. Which is now.
Legendary?
Now, why is MINIX considered legendary? Well, because MINIX, in combination with Andy Tanenbaum’s books on operating system design, was the blueprint for what later would become the biggest free and open source operating system of the world– yes, Linux. In Linus Torvalds’ autobiography, “Just for Fun”, Linus says that Tanenbaum’s book “Operating Systems: Design and Implementation” and MINIX were what “launched me to new heights”. More on this here.
That book, co-authored with Albert S. Woodhull, explains the inner workings of the MINIX operating system, and as a result the MINIX source code was sold together with the book. “Operating Systems Design and Implementation“ has therefor also been revised, in order to reflect MINIX 3.0.
About 3.0
MINIX 3.0 is released under a BSD-like license, and can be freely downloaded, altered, and so forth. In contrary to the Linux kernel (monolithic) and the WinNT/OSX kernels (hybrid), MINIX is a microkernel operating system. This crucial difference between MINIX and Linux led to one of the most famous flamewars in computer history, between Torvalds and Tanenbaum, held in comp.os.minix. You can read an abstract here. As a result of MINIX being a microkernel, that part of the kernel that lives in kernelspace consists of only 3800 lines of code. All device drivers (except the clock) live in userspace.
Over 300 UNIX programs are available for MINIX 3. It is POSIX-compliant, available for x86 (ARM7 and PPC ports under way) and supports up to 4GB of memory. A port of X Windows is also underway.
On the website, it is made clear that MINIX 3.0 is by no means as complete and full-featured as BSD or Linux. It is also explained that besides the traditional education market, MINIX 3.0 is also aimed at the embedded market, and applications where the GPL is too restrictive.
–Thom Holwerda
The bzip/IDE download seems to be corrupted though.
Maybe not, I must’ve just had a bad download
Yeah, my browser download cut out suddenly with no warning at about 90%. Using wget now instead.
Is it compatible with something else?
The MINIX license [ http://www.minix3.org/license.html ] is (a slightly modified version of) the Modified BSD License [ http://www.gnu.org/licenses/license-list.html#ModifiedBSD ] (i.e. without the advertising clause), so it should be compatible with pretty much everything, including the GPL.
X Window System, you ignorant.
It’s noticed like that on Tanenbaum ‘s website
> Well, because MINIX, in combination with Andy Tanenbaum’s
> books on operating system design, was the blueprint for what
> later would become the biggest free and open source operating
> system of the world– yes, Linux
Linux was never based on Minix’s code. The author may want to
clarify this statement. We do not want another FUD wave like the
one ADTI started.
Linux was never based on Minix’s code.
I know. Where did I state otherwise? “Was a blueprint” does not mean “based on”. Fact remains that MINIX and the book served a great and important inspiration for Linus. And that should not be left unmentioned.
More info on that here:
http://www.minix3.org/doc/faq.html#legal
I would even go as far as saying that “bluprint” is an overstatement. Not a small one at that.
“I would even go as far as saying that “bluprint” is an overstatement. Not a small one at that.”
…and you would still be wrong. Do your homework before posting stupidities.
If anything could be called a blueprint for Linux it would be the POSIX standard.
I’ve done my homework by reading “Linus Torvalds – Just for fun” about two years ago and I just flipped through it again. What homework have you done?
I think blueprint is an overstatement.
Minix gave some teaching and inspiration to the programmers of Linux may be towards a more accurate statement.
I seldom reply to these things, but when I saw the word “blueprint” I thought the exact same thing. Google “define:blueprint” and you will see words like “plan”, “guide”, “design”, and “template”. Sounds like it means “based on” to me! “Inspiration”, on the other hand, seems like an appropriate term, and my suggestion would be to update the article accordingly.
By the way, I have read both “Just for Fun” (Torvalds book as well as book on CD) and Tanenbaum’s OS book, 2nd edition. I was also inspired by the book to write my own little OS kernel, but Minix was certainly not a blueprint. So before anybody accuses me of not “doing my homework” or some nonsense, be aware of that.
> Linux was never based on Minix’s code. The author may want to
> clarify this statement. We do not want another FUD wave like the
> one ADTI started.
Ahh yes FUD. Thanks anonymous for actually creating FUD where it did not exist in its meaning.
I don’t see the big deal about Thom’s statement. Linux starting using Minix, and its modification became Linux. Thus, there’s no doubt Minix was the launching vehicle for Linux. Heck, the first Linux filesystem was the Minix filesystem. Having said that, Linux became something else at the end of all, and they are both different animals. But if Minix had not existed, possibly we would not have Linux today, and credit should be due where is due!
damn that’s an ugly logo
…to my uni days. Used to spend ages playing with the MINIX source. Might have to download for old times’ sake.
I think its better if Minix, for desktop users, does have a DirectFB implementation under the Desktop Environment, instead of X server.
Obligatory begging for some screenies!
Wanna have a screenie of a shell prompt? 😀
Obligatory begging for some screenies!
In Windows:
Start => Run… => “cmd.exe” => Full screen.
In Linux/GNOME:
Applications => System tools => Terminal => Full screen.
In Linux/KDE:
K => System tools => Terminal => Full screen.
In OSX Tiger:
[Apple]+[Space] => “Terminal” => Select Terminal from results => Full screen.
There, screenshots of MINIX
.
OMG TORRENT PLS!!!1122
Ah.. I used Minix to do my C programing at home during college.
In fact, I started using Linux in 1994 because of the buzz in the Minix community about it….funny to think that many people think an OS I have been using for more than a decade now is “NEW”.
Since some people have such a poor imagination.. Screenshots!
http://80.237.144.96/~christian/minix_311-screens/
slick~n~glossy!
I was a bit surprised that Minix is still being developed. But this is good since Minix has been perceived to be the practical starting ground for microkernels.
However looking at the Minix website I miss some documentation regarding Minix’s real-time capabilities. Does Minix do _hard_ real-time? And what is it’s basic scheduling mechanisms?
If it’s capable of doing hard real-time this would be great news for the embedded world for many purposes.
Linux in its current shape will not make it in embedded systems due to its lack of support for HRT. Maybe Minix is the way to go?
What is needed today is an implementation of a microkernel which implements HRT support and also support all the wonderfull applications of the GNU/Linux world. This would rock both the desktop/server world and the embedded world for sure.
I remember reading that Timesys had hard realtime patches for the Linux kernel at 2.4. When I was researching RTOSs they sent me a bunch of documentation on it and explanations about the architecture. It was pretty interesting, though I never actually downloaded their patches and looked at them. It was a good overview of what real-time systems do though.
> What is needed today is an implementation of a microkernel which implements HRT support and also support all the wonderfull applications of the GNU/Linux world. This would rock both the desktop/server world and the embedded world for sure.
What is the point of a real-time OS for desktop or server use? Desktop and server users want throughput, they do not care about real-time…
Anyway, this is great news but unfortunately the website is 503 for the moment…
Seems Tanenbaum will provide extensive documentation in the 900pp. book on Operating Systems concepts which uses Minix3 as an example, and comes out 5 days before Christmas.
Hmm, I wonder if I can afford the hobby…
Had you released this 15 years back, Linus would not have come out with Linux. Linux is usable, now why do we want a great educational OS to be practically usable? Stick to the goal of a good text book OS please.
“Linus Torvalds, then a student at the University of Helsinki, studied MINIX in an operating systems course and was sufficiently impressed that he bought a PC to run it.“
So, that implies he was running the Minix code in his mind by reading the source before he decided to run it on a PC? I wonder how many Bogomips he was getting before he got his PC to run it faster for him?
BTW, I helped test Minix3 for throughout the Summer on VMware, and I have to say that it runs awesome on it. It’s actually a breeze to install and run on VMware and pretty much any other platform.
After the basic install, which is way easier than before (no diskettes needed), you end up with a workable system. In the case of VMWare, you won’t need to even enable (DHCP) networking, as it is up and running by default.
Andy Tanenbaum was working with a few fulltime coders, particular Ben Gras, who did an incredible job.
For example, the driver for the network card on VMware had not been ported/updated for Minix3, and they were nice enough to help me (and everyone else at this point) with that.
A couple of biggies with regards to Minix3 are the following:
-The install is easier than ever before, and faster too. There’s no need to deal with ‘part’ (similar to fdisk, but kind of more complicated)
-Although it still has the requirement of at least two main partitions (root [boot monitor] and user), there’s no need to hit ‘=’ or anything else while booting. The exception is the ability to choose memory size before booting.
-cleaner code and many system updates (checkout the book when it comes out)
-gcc: It had been ported years ago, but it was not working; now it is.
-More updated network nic drivers than ever.
-A bunch of applications were ported and work from the start. Note that Minix does not have the ‘less’ command, as ‘more’ in Minix does the same thing.
My only complain is that I would love to see the Boot Monitor to go away for good, but that’s something small in whole deal…
What’s the deal with the 75 MB 3.1 .iso (link dead) on the previous versions versus the 10 MB 3.1.1 .iso (link active) on the download page? Was the 75 MB image the one from the book?
http://www.minix3.org/previous-versions/index.html
http://www.minix3.org/download/
I think with earlier releases for a small group of testers that no effort was made to zero out unused blocks in the CD image.
Taken horribly out of context but interesting to note.
From: [email protected] (Linus Benedict Torvalds)
Subject: Re: LINUX is obsolete
Date: 29 Jan 92 23:14:26 GMT
Organization: University of Helsinki
“True, linux is monolithic, and I agree that microkernels are nicer. With
a less argumentative subject, I’d probably have agreed with most of what
you said. From a theoretical (and aesthetical) standpoint linux looses.
If the GNU kernel had been ready last spring, I’d not have bothered to
even start my project: the fact is that it wasn’t and still isn’t.”
From:
http://www.oreilly.com/catalog/opensources/book/appa.html