I haven't been very active in this newsgroup but I do thank those who have been for keeping MINIX going. Some of the many people who have helped tremendously in the past are Kees Bot, Philip Homburg, Al Woodhull, Claudio Tantignone, Michael Temari, and, Giovanni Falzoni, but there are more. My thanks to all of you.
My group is now working on a new and much improved MINIX 3. This is a serious effort, with two experienced, professional programmers working full time on it now, not to mention various students. The goal for the first release is the end of October 2005. The main goals of MINIX 3 are to be small, reliable and secure. Numerous other improvements are in the works including longer file names, bigger memories and disks, more real-time friendly, better security, etc.
I would appreciate your not publicizing this now. It is far better to wait until there is a solid product available for downloading. Having a lot of people judging a pre-alpha release as the final product will be very bad PR.
I see a serious 'market' for MINIX 3 in several areas, among them:
1. Education, as has always been the case. The book is being updated.
2. Low-end PCs. Various organizations are working on PCs that will sell for under $100 in India and China and will be powered by batteries recharged by a crank, solar cells, or other sources of power where there is no electricity. An operating system for these limited computers must be small, resource efficient, and reliable.
3. Embedded systems such as DVD players, digital cameras and camcorders, TV sets, cell phones, and the like often have an operating system with multiprogramming and a hierarchical file system. They often need to be highly modular. Some are real time.
4, Companies who want a small (real-time) modular, open source operating system free of the GPL. MINIX is and will continue to be released under the BSD license.
I am looking for volunteers to help out. Areas where help is wanted include:
1. Porting software. A list of programs that we already have (or are working on) is given below. Other programs are welcome. See below for a few suggestions.
2. Porting drivers. Examples are - disk drivers (IDE with DMA, S-ATA, SCSI) - Ethernet drivers other than RealTek and Intel Pro/100 - USB and FireWire drivers, especially USB CD-ROMs - Printer drivers of all kinds - Audio drivers, - mouse drivers - etc., etc.
3. Porting MINIX 3 to hardware platforms other than the Pentium. Some work is underway for the PowerPC and ARM7, but others are welcome.
4. Porting new file systems. Since a file system is just a user program, multiple file systems can run at the same time. The Linux Ext3 file system is an obvious candidate. The CP/M file system would be fine for a digital camera.
5. More and tougher test suites (POSIX and application programs).
6. Testing MINIX 3 on various hardware configurations.
Other suggestions are welcome. Since we are now finalizing the OS itself for the book, we are NOT looking for help the the OS itself right now.
A prototype for the new Website is at www.minix.net. This is about MINIX 2, but shows you the new style. Comments are welcome.
We are also looking for a name and a logo. Current OSes have penguins, longhorns, tigers, daemons, and whatnot as symbols. Maybe we need an animate symbol too. Something that suggests small, tough, reliable, might be nice. I thought of the cockroach, but it got shot down internally because although they are tough as nails and have survived 300 million years, they are still bugs. And they are not cute. Maybe a female 'mailman' instead?
Comments to me and thanks for braving this long period of inactivity.
--------------------------------------------------------------------------- -- Here are the programs we already have or are working on:
Makefile comm finger isoread nice rm termcap aal compress flex join nl rmdir test add_route copy fmt kermit nm roff tget advent cp fold kill nohup rsh time ascii crc fortune last nonamed scripts touch ash cron fsck lbracket od sed tr at csplit fsck1 leave passwd seq traverse atrun cut ftp life paste setuidgid treecmp autil date ftpd link patch sh true awk dd gather ln pathchk sha1sum tsort backup de gcc login perl shar ttt badblocks decomp16 getty logname pine shred tty banner df gomoku look ping simple umount basename dhrystone grep lp pinky size uname bash diff groff lpd pr sleep unexpand bc dircolors gunzip ls pr_routes sort uniq bison dirname gzip m4 prep split unlink btoa dis88 head mail pretty stat unshar byacc diskcheck host make printenv strings update cal diskusage hostaddr man printf strip uptime calendar du hostid md5 printroot stty users cat dw hostname md5sum proto su uud cawf echo i386 men ptx sum uue cd ed i86 mesg pwd swapfs vol cdiff eject ibm mined pwdauth sync wc cgrep elle ic mkdir rarpd synctree whatsnew chgrp elvis id mkfifo rcp tac which chmem emacs ifconfig mkfs readall tail who chmod env ifdef mknod readfs talk whoami chown expand in.fingerd mkproto readlink talkd width chroot expr in.rld modem reboot tar write ci factor in.rshd mount recover tcpd xargs cksum false indent mref remove tee X11 clr fgrep inodes mt remsync telnet yap cmp file install mv rev telnetd yes co find irdpd ncheck rlogin term zmodem
Some programs for which we don't have any version or only an old version are listed below. Some require X11, which we don't have yet, so these should probably wait until X11 is ported. These are only examples. Anything useful is welcome. Here is the list:
Major libraries (e.g., zlib, pthreads, libpcap, libpng) are also welcome Test suites are very welcome, including tests of TCP (synfloods, bad packets), POSIX tests, and stress tests that push the system very hard.
Wow, I didn't see _this_ coming. Nice work. One question, though. In order to help out with porting and such, wouldn't we need to download the system-in-progress?
Also, most of the links on the new page are broken. And, well, the new design is, uhm, maybe a bit '70s. :)
* Andy Tanenbaum wrote:
: | We are also looking for a name and a logo. Current OSes have penguins, | longhorns, tigers, daemons, and whatnot as symbols. Maybe we need an | animate symbol too. Something that suggests small, tough, reliable, might | be nice. I thought of the cockroach, but it got shot down internally because | although they are tough as nails and have survived 300 million years, they | are still bugs. And they are not cute. Maybe a female 'mailman' instead?
Why don't we just stick a pair of eyes and a tail on the M and call it an animal.
> A prototype for the new Website is at www.minix.net. This is about > MINIX 2, > but shows you the new style. Comments are welcome.
> We are also looking for a name and a logo. Current OSes have penguins, > longhorns, tigers, daemons, and whatnot as symbols. Maybe we need an > animate symbol too. Something that suggests small, tough, reliable, > might be nice. I thought of the cockroach, but it got shot down > internally because although they are tough as nails and have survived > 300 million years, they are still bugs. And they are not cute. Maybe > a female 'mailman' instead?
> Comments to me, and thanks for braving this long period of inactivity.
I think the canonical image of a message passing system is two high-school girls at their desks, exchanging a folded slip of paper :-). Maybe just the hands and the note?
I hope my posted up-date to 'diff' is included. I can offer an updated 'man' page for the new diff options, too.
On Sun, 26 Jun 2005, Stian Sletner wrote: > Also, most of the links on the new page are broken. And, well, the new > design is, uhm, maybe a bit '70s. :)
Hey, I -loved- the web page designs in the '70s.
> * Andy Tanenbaum wrote:
> > We are also looking for a name and a logo. > Why don't we just stick a pair of eyes and a tail on the M and call it > an animal.
At one time I tried to draw something like the big (mechanical?) spider from Johnny Quest out of the M. It didn't work well but it might have been lack of talent...
Other possibilities, (several previously suggested): A (draft) horse, w/ collar, kinda (but not too much!) like Horace. An Armadillo A bull dog (Warneresqe. --Big up front, tiny backside. Spiked collar) A mechanical mouse w. a big wind-up key. A mouse w/ the M on his shirt. A polar Bear (south->penguins, north->polar bears) Other insects have been suggested, like a ladybug or an ant A worm or caterpillar (maybe not a worm) A Frog An alligator A wombat/mole/vole/unidentified but small & furry. Pegasus or other mythical critter/charactor An angel (calm those anti-bsd daemon types) A witch (further upset those abdt) A duck Scales, a duck & a witch (get the monty python fans) Alien or bug-eyed monster A robot (named Isaac?)
Note that slackware linux has a person. Maybe a Dr. Tee? Oliver Wendell Jones and/or the banana 9000 (hey, you can ask,)
Or maybe even Veronica Karlsson's ascii Kitten (with permission of course):
On Sat, 25 Jun 2005, Andy Tanenbaum wrote: > We are also looking for a name and a logo. ... > Maybe a female 'mailman' instead?
Oh, wait. Like the mailman on the OSDI cover? Yeah, he'd make a good mascot. Your gonna make a lady out of him, huh?
I also like the robot w/ the box of mail on Computer networks, 3rd edition too.
Or a pigeon w/ a mail bag like on "stop that pigeon" http://www.hotink.com/wacky/dastrdly/bird2.gif (I'd mix in a good dollop of Walt Kelly's Mallard De Mer in the drawing --though I can't find an online pic)
> 2. Porting drivers. Examples are > - disk drivers (IDE with DMA, S-ATA, SCSI)
It might be useful to investigate compatibility frameworks that would allow other OS' drivers to be used; for example, the BSD network driver model is very clean and efficient (and their drivers tend to be extremely well written). Being able to link one of their drivers with a library and then use the result as a Minix driver would be a huge bonus.
[...]
> 3. Porting MINIX 3 to hardware platforms other than the Pentium. Some > work is underway for the PowerPC and ARM7, but others are welcome.
I have an 8MB-RAM MMUless ARM7 device that I'd be interested in doing a port to; Minix would make a great alternative to ucLinux, which I'm not very keen on.
Incidentally, will Minix 3 still support 16-bit systems?
[...]
> Other suggestions are welcome. Since we are now finalizing the OS itself > for the book, we are NOT looking for help the the OS itself right now.
It would be a fairly large project, but you might want to consider building a Debian platform around the Minix kernel. Debian isn't limited to Linux; there are versions based around the HURD, NetBSD and FreeBSD kernels. This would certainly require a 32-bit Minix, and probably require a MMU, but it would solve the whole software bootstrapping problem at a stroke, and provide obvious proof of capability.
> A prototype for the new Website is at www.minix.net. This is about MINIX > 2, but shows you the new style. Comments are welcome.
I like it. Clean and well-organised. The colour scheme is a little 80s, but that's not necessarily a drawback.
-- +- David Given --McQ-+ "`Aplysia californica' is your taxonomic | d...@cowlark.com | nomenclature. | (d...@tao-group.com) | A slug, by any other name, is still a slug by +- www.cowlark.com --+ nature." --- drushel on a.f.c
An ant would make a nice logo. It would symbolise the collective work that an open source OS requires. The ant is also powerful; it can carry about 30 times its weight
> We are also looking for a name and a logo. Current OSes have penguins, > longhorns, tigers, daemons, and whatnot as symbols. Maybe we need an > animate symbol too. Something that suggests small, tough, reliable, might > be nice. I thought of the cockroach, but it got shot down internally because > although they are tough as nails and have survived 300 million years, they > are still bugs. And they are not cute. Maybe a female 'mailman' instead?
Is MINIX 3 evolutionary or revolutionary? In other words is it based on the code of MINIX 2 with modifications and extensions or is it a total rewrite from the ground up?
I believe that apart from the porting and the development process for release 3, the minix community should also start a process for regulating and categorizing different software ports to the minix platform. Although, this may at first sound like impeding the growth of the platform, I think this is a tremendous opportunity for minix to score over Linux and other so called "Open source" systems. The reason why I suggest this is twofold:
>Since minix is largely used by academic community, porting and including everything will eventually lead to a
titanic distribution - the way it has happened with Fedor/Red hat and almost all other linux distributions. Trust me, most users are not even aware of what those 6 installation CD's contain. Such huge distributions overwhelm the user at first and restrict the growth. Most students dont really need that much software.
>Almost all users dont need everything. A better option would be to provide the core OS as one unit and other
software as optional different pacakges/bundles, according to their respective functions e.g- Admin package, Developer package etc.
I think since minix is only in its nascent stages, the earlier such classification is adopted, the better.
Comments are welcome.
Further, I also express my will to port Minix to AMD 64 bit platform, since I have one such machine. Although, I would initially be porting the core, I would be greatful if someone could provide me with some leads about device drivers, esp. sound and video. I already have the linux source code for my display card so porting should not be such a big issue.
Anyways, helps and comments are welcome.
I would like to reiterate my willingness to do all the dirty work i.e. porting device drivers and writing low level code for amd64 port, and request anyone who may provide some leads to please do so.
I think Sanket's suggestion to not overwhelm everyone with masses of software is a good one. I can envision the core distribution being one CD with other packages on the website.
I am definitely interested in a port to the AMD. A port to the PowerPC is underway now as well as the ARM7.
The book schedule is now firming up. The release will be frozen Sept 2. For this reason, heavy testing of 3.0.6 and bug reporting is much appreciated. When 3.0.7 is available, I will post a notice here.
Is Minix 3 coming up with a Window X interface or with virtual memory, just like Minix-VMD? I mention the two characteristics still required by Minix to be considered a full fledged Unix-like OS.
David Given <d...@cowlark.com> wrote: > Andy Tanenbaum wrote: > [...] > > 2. Porting drivers. Examples are > > - disk drivers (IDE with DMA, S-ATA, SCSI) > It might be useful to investigate compatibility frameworks that would allow > other OS' drivers to be used; for example, the BSD network driver model is > very clean and efficient (and their drivers tend to be extremely well > written). Being able to link one of their drivers with a library and then > use the result as a Minix driver would be a huge bonus. > [...] > > 3. Porting MINIX 3 to hardware platforms other than the Pentium. Some > > work is underway for the PowerPC and ARM7, but others are welcome. > I have an 8MB-RAM MMUless ARM7 device that I'd be interested in doing a port > to; Minix would make a great alternative to ucLinux, which I'm not very > keen on. > Incidentally, will Minix 3 still support 16-bit systems? > [...] > > Other suggestions are welcome. Since we are now finalizing the OS itself > > for the book, we are NOT looking for help the the OS itself right now. > It would be a fairly large project, but you might want to consider building > a Debian platform around the Minix kernel. Debian isn't limited to Linux; > there are versions based around the HURD, NetBSD and FreeBSD kernels. This > would certainly require a 32-bit Minix, and probably require a MMU, but it > would solve the whole software bootstrapping problem at a stroke, and > provide obvious proof of capability. > > A prototype for the new Website is at www.minix.net. This is about MINIX > > 2, but shows you the new style. Comments are welcome. > I like it. Clean and well-organised. The colour scheme is a little 80s, but > that's not necessarily a drawback. > -- > +- David Given --McQ-+ "`Aplysia californica' is your taxonomic > | d...@cowlark.com | nomenclature. > | (d...@tao-group.com) | A slug, by any other name, is still a slug by > +- www.cowlark.com --+ nature." --- drushel on a.f.c
