davem@sdcrdcf.UUCP (David Melman) (01/21/87)
I'm curious how the goals and purpose of Comer's Xinu operating system (Operating System Design, the Xinu Approach) differ from Tanenbaum's Minix. A few things they do have in common are: 1) They both run on micros 2) They both are a rewritten simplification of Unix 3) They both are intended to be educational David Melman
mcvoy@rsch.WISC.EDU (Lawrence W. McVoy) (01/21/87)
In article <3862@sdcrdcf.UUCP> davem@sdcrdcf.UUCP (David Melman) writes: > >I'm curious how the goals and purpose of Comer's Xinu operating system >(Operating System Design, the Xinu Approach) differ from Tanenbaum's Minix. > >A few things they do have in common are: > 1) They both run on micros I guess, if you call an LSI-11 a micro. > 2) They both are a rewritten simplification of Unix Nope. Xinu stands for "Xinu Is Not Unix". Did you read the book? Hmm... A long time ago, maybe? Oh, well. > 3) They both are intended to be educational I disagree. Let me take a stab at the answer to your first question. I think that Xinu is educational. It may not have been intended as such, but because of the lack of supporting software & hardware it will never amount much more than an intellectual exercise. (Don't get me wrong -- I got a lot out of the Xinu book). Minix is an attempt to remove all those footnotes from journal articles that say "Unix is a trademark of Ma Bell". It is also an attempt to standardize a Unix so that software developers have something to shoot for. It's also a replacement for DOS, no small item when you consider there are probably (note the clause) more IBM-PC's & clones out there than any other single CPU. A large potential market. Last but not least, Tanenbaum stands to make a fair bit off his OS book. He already has a pretty good reputation for cranking out readable texts. -- Larry McVoy mcvoy@rsch.wisc.edu, {seismo, topaz, harvard, ihnp4, etc}!uwvax!mcvoy "They're coming soon! Quad-stated guru-gates!"
ralphw@ius2.cs.cmu.edu (Ralph Hyre) (01/22/87)
In article <3862@sdcrdcf.UUCP> davem@sdcrdcf.UUCP (David Melman) writes: > >I'm curious how the goals and purpose of Comer's Xinu operating system >(Operating System Design, the Xinu Approach) differ from Tanenbaum's Minix. > >A few things they do have in common are: > 1) They both run on micros > 2) They both are a rewritten simplification of Unix > 3) They both are intended to be educational And they're both 'published' by Prentice-Hall. You apparently need a UNIX source license to get a XINU distribution tape from Prentice-Hall. This is a diasadvantage. It will be interesting to see if the publisher will continue to support a product with such restrictions. Here is an excerpt from the article I saw that mentioned XINU. ]Date: Wed, 20 Nov 1985 18:16-EST ]From: th@python.UUCP (T. Hartnett) ]To: ralphw@ius2.cs.cmu.edu.ARPA ]Subject: A vote for net.os (and some XINU) ]Message-ID: <net.unix/4328> .... ] If you are working with an LSI-11, most everything is done for you ]already. You just have to get a hold of the distribution tape from Prentice- ]Hall (it's approx. $60 plus a copy of your UNIX license). -- - Ralph W. Hyre, Jr. Internet: ralphw@ius2.cs.cmu.edu Phone: (412) 268-2847 [CMU-BUGS], 268-3275 Amateur Radio: KA3PLY temporary KT (c/o W3VC, CMU Radio Club, Pittsburgh, PA)
tsf@theory.cs.cmu.edu (Timothy Freeman) (01/22/87)
>In article <3862@sdcrdcf.UUCP> davem@sdcrdcf.UUCP (David Melman) writes: >>I'm curious how the goals and purpose of Comer's Xinu operating system >>(Operating System Design, the Xinu Approach) differ from Tanenbaum's Minix. >>A few things they do have in common are: >> 1) They both run on micros >> 2) They both are a rewritten simplification of Unix >> 3) They both are intended to be educational I and several cohorts implemented a subset of Xinu for the IBM PC as a project for an Operating Systems course back when I was at Rensselaer. Here are some differences between the two. (I have no knowledge about Minix except by reading this bboard, so my statements about Xinu are going to be more accurate than my statements about Minix.) 4) Minix is System V compatible; Xinu isn't meant to be a Unix, so it isn't compatible with anything. (Xinu stands for Xinu Is Not Unix, and I can vouch for the correctness of its acronym.) 5) Xinu only has lightweight processes (that is, they are all in the same address space and share a common pool of resources). Minix only has heavyweight processes (that is, each process has its own address space and can try to believe that it has its own pool of resources, to the extent to which this is true for Unices in general.) 6) Xinu doesn't have an analogue to the exec system call, all code that you can ever execute must be in the initial core image. Minix does have an exec call. 7) Xinu specifies a new process to start by giving a subroutine to execute for the new process; it doesn't have a fork call. Minix does have fork. 8) Xinu was not sufficient to use as a development system for itself. Minix hopefully is. (The original Xinu development environment ran under Unix and used cross-compilers. The Xinu version I developed used MS-Dos as a development environment.) In summary, Xinu is a great vehicle for an operating systems course because it is possible for one person to implement enough of it to be able to run things within a semester, but don't try to use it to get any work done. -- Tim Freeman Arpanet: tsf@theory.cs.cmu.edu Uucp: ...!seismo!theory.cs.cmu.edu!tsf
brent%terra@Sun.COM (Brent Callaghan) (01/22/87)
In article <1053@ius2.cs.cmu.edu>, ralphw@ius2.cs.cmu.edu (Ralph Hyre) writes: > You apparently need a UNIX source license to get a XINU distribution > tape from Prentice-Hall. This is a diasadvantage. Huh! If Xinu is not Unix then why the Unix license ? (I'm confused) Made in New Zealand --> Brent Callaghan @ Sun Microsystems uucp: sun!bcallaghan phone: (415) 691 6188
budd@bu-cs.UUCP (01/23/87)
Xinu as described in the book was not an attempt at providing a user oriented operating system. For example there are no provisions for user address spaces (fork takes the address of a routine to call) The tape from the publisher contains a complete cross development environment for VAX/BSD, which includes a modified version of the Un*x V7 C compiler, loader etc in source code. I believe the only reason for the licence is to obtain the v7 tools.
mwm@cuuxb.UUCP (01/24/87)
In article <3862@sdcrdcf.UUCP> davem@sdcrdcf.UUCP (David Melman) writes: > >I'm curious how the goals and purpose of Comer's Xinu operating system >(Operating System Design, the Xinu Approach) differ from Tanenbaum's Minix. > >A few things they do have in common are: > 1) They both run on micros > 2) They both are a rewritten simplification of Unix > 3) They both are intended to be educational > >David Melman Wrong on 2). Xinu is NOT Unix. (X.I.N.U. -- get it?) Xinu is a smaller, simpler, multitasking operating system used to teach operating system concepts. A Unix-like interface could be layered onto Xinu, but it has not (to my knowledge) been done. So you cannot take a random Unix program and compile the code and run it under Xinu. It would seem to me that one might want, when teaching about operating systems, want to go through both systems -- Comer's because it is easy to understand and play with, and then Tannenbaum's because it provides many features that Xinu does not. (for example, Xinu provides no multi-user facilities like user-id's, file ownership, etc.) -- Marc Mengel ...!ihnp4!cuuxb!mwm
andrewjp@uoregon.UUCP (01/24/87)
>You apparently need a UNIX source license to get a XINU distribution >tape from Prentice-Hall. This is a diasadvantage. But then, since XINU is in the public domain, you don't really need a copy of the distributon tape. We have xinu installed at our site, and I'm sure many others have it lying around as well. There's more than one way to skin a walrus. Cygnus. "Better Living Through Ignorance" Eric Swanson c/o andrewjp@drizzle.UUCP P.O. Box 30098 Eugene, OR 97403 (503) 484-2790 or (503) 484-4184
john@moncol.UUCP (01/25/87)
In article <1053@ius2.cs.cmu.edu> ralphw@ius2.cs.cmu.edu.UUCP writes: >In article <3862@sdcrdcf.UUCP> davem@sdcrdcf.UUCP (David Melman) writes: >> >>I'm curious how the goals and purpose of Comer's Xinu operating system >>(Operating System Design, the Xinu Approach) differ from Tanenbaum's Minix. >> >>A few things they do have in common are: >> 1) They both run on micros >> 2) They both are a rewritten simplification of Unix >> 3) They both are intended to be educational > >And they're both 'published' by Prentice-Hall. >You apparently need a UNIX source license to get a XINU distribution >tape from Prentice-Hall. This is a diasadvantage. >It will be interesting to see if the publisher will continue to support >a product with such restrictions. Just a slight clarification... the problem is not that you need a UNIX license to get the XINU sources, but to get the distribution from Prentice-Hall which includes some UNIX-derived cross-development tools. Some time ago, someone actually posted (or offered to mail) the XINU sources. He had permission from Comer and there was no legal hassle since he was not distributing the cross development stuff. -- Name: John Ruschmeyer US Mail: Monmouth College, W. Long Branch, NJ 07764 Phone: (201) 571-3557 UUCP: ...!vax135!petsd!moncol!john ...!princeton!moncol!john ...!pesnta!moncol!john If you give bigger kisses, you get bigger hugs.
ron@brl-sem.UUCP (01/26/87)
In article <3862@sdcrdcf.UUCP>, davem@sdcrdcf.UUCP (David Melman) writes: > > I'm curious how the goals and purpose of Comer's Xinu operating system > (Operating System Design, the Xinu Approach) differ from Tanenbaum's Minix. > > A few things they do have in common are: > 1) They both run on micros > 2) They both are a rewritten simplification of Unix > 3) They both are intended to be educational > I'll add a few more 4) The source to both are available from Prentice-Hall. The main advantage to MINIX, I see, is that it comes ready to fly and experiment on an IBM-PC, which most people have or can get a hold of. XINU is for the PDP-11, which are becoming scarce, and have never been dirt cheap. Also, I believe, that you really have to have a PDP-11 development environment (read C compiler) to run it, while MINIX has the binaries for the PC one on the disks. -Ron
ron@brl-sem.UUCP (01/26/87)
In article <1010@theory.cs.cmu.edu>, tsf@theory.cs.cmu.edu (Timothy Freeman) writes: > > 4) Minix is System V compatible; Xinu isn't meant to be a Unix, so > it isn't compatible with anything. (Xinu stands > for Xinu Is Not Unix, and I can vouch for the correctness of > its acronym.) MINIX IS NOT SYSTEM V compatible, but it does bear a striking resemblence to large parts of Version 7 UNIX. XINU is UNIX spelled backwards and any attempt to words to the acronym is an afterthought.
allbery@ncoast.UUCP (01/27/87)
As quoted from <11890@sun.uucp> by brent%terra@Sun.COM (Brent Callaghan): +--------------- | In article <1053@ius2.cs.cmu.edu>, ralphw@ius2.cs.cmu.edu (Ralph Hyre) writes: | > You apparently need a UNIX source license to get a XINU distribution | > tape from Prentice-Hall. This is a diasadvantage. | | Huh! If Xinu is not Unix then why the Unix license ? | | (I'm confused) +--------------- Because the distribution tape includes a XINU development environment: source for modified versions of pcc, as, and ld, among other things. Since they are derived from AT&T UNIX source, you need an AT&T UNIX source license to get them. And because it's so minimal as distributed, XINU is worthless without them (remember, XINU is a bare-bones OS intended to teach basic OS concepts). ++Brandon -- ____ ______________ / \ / __ __ __ \ Brandon S. Allbery <backbone>!ncoast!allbery ___ | /__> / \ / \ aXcess Co., Consulting ncoast!allbery@Case.CSNET / \ | | `--, `--, 6615 Center St. #A1-105 (...@relay.CS.NET) | | \__/ \__/ \__/ Mentor, OH 44060-4101 \____/ \______________/ +1 216 781 6201
john@frog.UUCP (01/28/87)
> > You apparently need a UNIX source license to get a XINU distribution > > tape from Prentice-Hall. This is a diasadvantage. > Huh! If Xinu is not Unix then why the Unix license ? > (I'm confused) > The XINU distribution tape from Prentice-Hall includes some modified UNIX utilities (a standalone C compiler, linker, and library, I believe), and rather than have a tape with and a tape without, they chose to always require a license (perhaps because their lawyers weren't used to this kind of problem, perhaps because they were lazy). -- John Woods, Charles River Data Systems, Framingham MA, (617) 626-1101 ...!decvax!frog!john, ...!mit-eddie!jfw, jfw%mit-ccc@MIT-XX.ARPA WITH YOUR MESSAGE ***
jose@inmet.UUCP (01/29/87)
It is unfortunate that people view XINU as a simple operating system for educational purposes. It is in fact a viable alternative for developing software on bare machines. Unix (TM) and MINIX are not well suited for use in embedded applications at this time. They are too big and offer many more facillities than are required. In contrast, XINU's size and process control facilities make it, in my opinion, as attractive an alternative for multi process embedded applications as Ada (TM) or Modula 2. Jose Oglesby Intermetrics Inc.
awol@vpnet.chi.il.us (Al Oomens) (02/09/91)
I would appreciate any info/opinions regarding differances/advantages of MINIX over XINU and vice-versa. If you would mail replys, I would be willing to sumarize. Being completely new to both OSs, I am considering implementing one or the other, and would appreciate any info. Thanks!