ast@cs.vu.nl (Andy Tanenbaum) (07/07/89)
I have been fairly quiet lately, not out of lack of interest, but waiting for the dust to settle. It seems that the overwhelming majority opinion is that Bruce Evan's protected mode code is a big step forward. I tried it and, of course, it didn't work on my machine. However, after a round of 12,000 mile remote debugging with Bruce, we found the problem (2 lines) and it works fine now. I have also looked at his code more closely, and I am impressed. It is not only very well programmed, but he resisted the temptation that most other people have succumbed to, namely, changing my layout. If I am to use it in the book, I certainly don't want two different styles in there, and my changing everything by hand would undoubtedly lead to errors. There is no MINIX pretty-printer at present (hint). The UNIX pretty-printer, cb, doesn't produce my layout style, which I am very much attached to. Another plus is that Bruce has been very helpful in answering questions etc. My major fear, that the code would be too complicated for students, is probably unjustified. It is certainly longer now, but not that much hairier. The 80286 stuff is fairly well isolated. I took a look at how long the listing in the book will be if I include only the same files as last time, plus a few new ones that are essential. It comes to something like 380 pages, vs. 250 last time. Together with, say, 30 pages of descriptive text about how the new code works, and 20 pages of man pages in Appendix C, the book will exceed 900 pages. This is getting kinda large for a 1 semester textbook. Printing it in 2 volumes makes it more expensive. I might be able to help a little here and there by not including things like klib88.x and protect.c, which are very technical and not very conceptual, and perhaps eliminating the cross reference listing. Anyway, I have decided to use Bruce's code as the basis for V2.0 (subject to Prentice-Hall not vetoing a 900 page book). There are still a number of things that need to be done before I even start with POSIX. Bruce and I have discussed this by email. The game plan is that he is going to make some changes during the summer, and then post the new version of the kernel. I am going to look at fixes etc from the net concerning utilities etc, and post those cdifs. All this should happen around September. Let us call that version 1.4b. This will be my base for starting to work on POSIX. I'll do the kernel/FS/MM myself (looking at what Simon Poole has already done), but help writing/converting libraries, utilities etc is very welcome. When all that is done, we will have a protected mode system that is based on POSIX and reasonably fast for small machines. MINIX seems to be evolving in the direction of something more than an academic toy and towards a more serious system. If I run into Niklaus Wirth at a conference sometime, I'll have to ask him about that :-). An additional note. Amiga MINIX is actually up and running. It is now being tested. Furthermore, MINIX 1.3 is being ported to the SPARC. Getting all these versions back together, and then POSIX-based is going to be a job on the order of cleaning out the Augean Stables, and I don't have a Hercules card. Andy Tanenbaum (ast@cs.vu.nl)
phil@june.cs.washington.edu (Phil Nelson) (07/08/89)
I don't have a 286 and so I didn't get a copy of Bruce's protected mode kernel. With the announcement that V2.0 will be based on Bruce's work, I had a question that was not answered by the announcement. What does this move mean to those of us without a 286? Will the same code run on both 8088s and 80286s? --Phil Nelson (aka phil@unicorn.wwu.edu)
jk0@image.soe.clarkson.edu (Jason Coughlin) (07/09/89)
From article <2835@ast.cs.vu.nl>, by ast@cs.vu.nl (Andy Tanenbaum): > ,and perhaps > eliminating the cross reference listing. I realize that with 900 pages it is hard to figure out what to cut and what not to cut. But, the cross reference listing is VERY handy. I used it ALL of the time when taking our OS course. If it's gotta go then it's gotta go, but if something else can go instead please leave it. -- -- Jason Coughlin ( jk0@sun.soe.clarkson.edu , jk0@clutx )
mshiels@tmsoft.uucp (Michael A. Shiels) (07/09/89)
Will the protected mode version work on a 386 machine? I am not familiar with the differences for protected mode. Thanx.
jca@pnet01.cts.com (John C. Archambeau) (07/10/89)
Might as well post this to everybody since this question will most likely be
asked again. Yes, the Evans' 286 protect mode kernel should run on a 386 or
386SX machine. Intel designed the 80x86 to be downward compatable. So if you
have a 386 or 386SX you should be able to run 8086 and 80286 code with NO
MODIFICATIONS whatsoever. Everything is there in the chip. The reason that
this works is that the unused bits in the 24-bit descriptor table are set to
zero, so when the 386 runs into a descriptor table with the high byte set to
zero, it's clicking in the back of its mind "ah ha! this person wants me to
emulate a brain damaged 286" and does so appropriately. I might be wrong of
where in the descriptor table the bits are set to zero, but I know all unused
bits are set to zero. See "The 80386/80387 Architecture" for details on 286
emulation on a 386. One of the authors is Issacson who is also the author of
the A86 assembler.
/*--------------------------------------------------------------------------*
* Flames: /dev/null (on my Minix partition)
*--------------------------------------------------------------------------*
* ARPA : crash!pnet01!jca@nosc.mil
* INET : jca@pnet01.cts.com
* UUCP : {nosc ucsd hplabs!hd-sdd}!crash!pnet01!jca
*--------------------------------------------------------------------------*/ast@cs.vu.nl (Andy Tanenbaum) (07/10/89)
In article <8678@june.cs.washington.edu> phil@june.cs.washington.edu (Phil Nelson) writes: >I don't have a 286. >What does this move [Bruce Evans' code] mean to those of us without a >286? Will the same code run on both 8088s and 80286s? Yes. Definitely. Bruce was extremely careful to distinguish where something was specific for the 286. In the next release, the test for 8088 vs. other CPUs will be dynamic. Conclusion: don't worry. We are not planning to dump the 8088 by any means. One thing I may do, however, is make the assumption that by the time this system (V2.0) is released, probably 1991, everybody wanting to recompile the system will have a hard disk. Making the system boot and run with floppy only isn't so hard, but as the kernel has grown, it is getting harder and harder to recompile the whole operating system on floppies. By 1991 hard disks will be so cheap that it is hard to imagine any serious user not having one. Andy Tanenbaum
ast@cs.vu.nl (Andy Tanenbaum) (07/10/89)
In article <1989Jul8.171336.27399@sun.soe.clarkson.edu> jk0@image.soe.clarkson.edu (Jason Coughlin) writes: >I realize that with 900 pages it is hard to figure out what to cut and >what not to cut. But, the cross reference listing is VERY handy. How about this. I could grep the listing in the book for PRIVATE and PUBLIC and just give an alphabetical list of those lines, 3 or 4 columns per page. That way you could find where symbols are defined, which I think is more useful than the reverse. Conceivably I could also make a full cross reference list and put in on the disk. Making that list, incidentally, was a gigantic pain in the rear end. It was surprisingly difficult to automate. Andy Tanenbaum
meulenbr@cstw01.prl.philips.nl (Frans Meulenbroeks) (07/10/89)
[lots of text discarded] In article <2835@ast.cs.vu.nl> ast@cs.vu.nl (Andy Tanenbaum) writes: [...] >Anyway, I have decided to use Bruce's code as the basis for V2.0 (subject to >Prentice-Hall not vetoing a 900 page book). [...] >starting to work on POSIX. I'll do the kernel/FS/MM myself (looking at >what Simon Poole has already done), but help writing/converting libraries, >utilities etc is very welcome. > >When all that is done, we will have a protected mode system that is based >on POSIX and reasonably fast for small machines. MINIX seems to be >evolving in the direction of something more than an academic toy and towards >a more serious system. If I run into Niklaus Wirth at a conference >sometime, I'll have to ask him about that :-). > >An additional note. Amiga MINIX is actually up and running. It is now >being tested. > >Furthermore, MINIX 1.3 is being ported to the SPARC. > >Getting all these versions back together, and then POSIX-based is going to be >a job on the order of cleaning out the Augean Stables, and I don't have a >Hercules card. > >Andy Tanenbaum (ast@cs.vu.nl) Ok. That amounts to 5 different versions: for 8088, 286/protected, ST, Amiga, SPARC. As far as getting these together: This can really become a pain. However, I would support such an integration very much. As far as I can see it there are two different problem areas: - assembly code Since these must be rewritten completely, IMHO they should be avoided as much as possible. I know that it is tempting to do a lot of rs232 stuff in assembly for performance purposes. However I would like to ask to limit this to the bare minimum. - hardware dependencies This is another problem, which mainly arises in the tty/rs232/console area. Please do not rely on the fact that every system has a 8250 chip. In the current tty.c there are a lot of places where the 8250 is enabled. I would suggest to keep all hardware independent stuff in tty.c, and either put all hardware stuff in say console.c and rs232.c or put it in dedicated files like i8250.c and so on, dealing with the various hardware components. Doing things this way yields a clear separation of the common part and the hardware specific part. If ast decides to isolate the hardware depencies in one way or another, I could be persuaded to look at the ST part of it. (By the way, the above does mainly apply to rs232, keyboard and monitor. I don't know if such an approach is feasible for floppy and hard disk). On a related but different topic: will something be done to get the compiler more ansi conformant? I'm mainly thinking about function prototypes and floating point support. As an aside: I have a version of 1.4a for the ST, but I really don't know what to do with it. Since I prefer very much to have an official upgrade I'm very reluctant to post this. On the other hand: I've shipped this to Johan Stevenson for authorisation and posting somewhere in April, but it seems that he does not have the time to look into it, and approve it. I'm sitting on this stuff for more than two months now, and I feel in some kind of catch-22 situation. What is the opinion of the net in general and ast specifically that I should do with this stuff?? Note: please don't send mail requesting me to mail you a copy. I want to get the minix community in sync, and I definitely don't want a lot of different version floating around. (by the way: this ST version is compatible with 1.4a, except for the amoeba stuff. It does support rs232 (at least at 1200 baud)) regards, Frans Meulenbroeks (meulenbr@cst.prl.philips.nl) Centre for Software Technology ( or try: ...!mcvax!phigate!prle!cst!meulenbr)
uri@arnor.UUCP (Uri Blumenthal) (07/10/89)
From article <1989Jul9.023852.21674@tmsoft.uucp>, by mshiels@tmsoft.uucp (Michael A. Shiels):
> Will the protected mode version work on a 386 machine?
I'm not familiar with MINIX. But if it works on 286 machine, it should
work on 386 AS WELL, since 386 can emulate 286. Details are not
necessary here and now, but I think it should work (but don't expect
all the 386 goodies!).
Uri.evans@ditsyda.oz (Bruce Evans) (07/11/89)
In article <1989Jul9.023852.21674@tmsoft.uucp> mshiels@tmsoft.UUCP (Michael A. Shiels) writes: >Will the protected mode version work on a 386 machine? I am not Yes, assuming the machine is AT-compatible apart from the processor. This does not use 32-bit mode so still suffers from small segments. It (the same binary) will also run on an 8088, by avoiding the small parts of the code specific to protected mode. I am working on a full 32-bit version. It runs, but there is no practical compiler for it yet. -- Bruce Evans evans@ditsyda.oz.au
charles@hpcvca.CV.HP.COM (Charles Brown) (07/11/89)
> This is getting kinda large for a 1 semester textbook. > Andy Tanenbaum (ast@cs.vu.nl) Just a nit... Why does the textbook need to be restricted to 1 semester? I have seen several textbooks which, in the introduction, map out various sizes (how many semesters) and levels (undergrad vs grad) of courses by suggestions on what chapters to cover and what chapters to skip. I think that approach works well. -- Charles Brown charles@cv.hp.com or charles%hpcvca@hplabs.hp.com or hplabs!hpcvca!charles or "Hey you!" Not representing my employer.
jk0@image.soe.clarkson.edu (Jason Coughlin) (07/11/89)
From article <2851@ast.cs.vu.nl>, by ast@cs.vu.nl (Andy Tanenbaum): > How about this. I could grep the listing in the book for PRIVATE and PUBLIC > and just give an alphabetical list of those lines, 3 or 4 columns per page. Sounds good to me. Thank you. -- Jason Coughlin ( jk0@sun.soe.clarkson.edu , jk0@clutx.BITNET ) -- -- Jason Coughlin ( jk0@sun.soe.clarkson.edu , jk0@clutx )
rbthomas@athos.rutgers.edu (Rick Thomas) (07/11/89)
> Conceivably I could also make a full cross > reference list and put in on the disk. Now, That's a good idea! Most people aren't interested in the Xref unless they are into serious code hacking anyway, so they can be expected to have the disks. Also, I bet it's easier to talk P-H into adding a disk to the distribution than adding 50-100 pages to the book. (Or printing a second volume with the code and Xref in it.) As an added benefit, if the Xref production can be automated (not easy, as Andy noted, but maybe some of the hackers out there would be willing to have a go at it. Warning -- you should probably start with the source code of the CPP and parser pass of the compiler. Anything less is bound to wind up getting confused by comments that look like code, etc.) The program that does it can be put on the floppy, instead of the actual Xref. Then people can have Xrefs of their own local versions. Rick -- Rick Thomas uucp: {ames, cbosgd, harvard, moss, seismo}!rutgers!jove.rutgers.edu!rbthomas arpa: rbthomas@JOVE.RUTGERS.EDU Phone: (201) 932-4301
ast@cs.vu.nl (Andy Tanenbaum) (07/11/89)
In article <563@prles2.UUCP> meulenbr@cstw01.prl.philips.nl (Frans Meulenbroeks) writes: >I have a version of 1.4a for the ST, but I really don't know what to do >with it. Since I prefer very much to have an official upgrade >I'm very reluctant to post this. >On the other hand: I've shipped this to Johan Stevenson for >authorisation and posting somewhere in April, but it seems that he does >not have the time to look into it, and approve it. I don't either. Still, I think it worth trying to get the two versions into sync. You could post the cdifs and let the net try to work out the bugs collectively. A far more interesting question is getting the second release of Bruce Evans' stuff and the Atari in sync, since that will be the base for V2.0 on the IBM line. Andy Tanenbaum (ast@cs.vu.nl)
ast@cs.vu.nl (Andy Tanenbaum) (07/11/89)
In article <Jul.11.01.40.45.1989.26685@athos.rutgers.edu> rbthomas@athos.rutgers.edu (Rick Thomas) writes: > Also, I bet it's easier to talk P-H into >adding a disk to the distribution than adding 50-100 pages to the >book. That's true, but next time around I will probably only have 360K diskettes. For V1.x, there was 640K PC and 512K AT. All the people with 512K PCs and 640K ATs got confused. Thus the number of diskettes will invariably be larger next time, which does not give any technical problems, but does affect the price. > The program that does it can be put on the floppy, instead of >the actual Xref. Then people can have Xrefs of their own local >versions. Not a bad idea, assuming I can automate the process. I'll also have to put the line numbering program on the disk. Not that it is hard to do that, but I might forget. Somebody please remind me around Dec 1990. Andy Tanenbaum (ast@cs.vu.nl)
ast@cs.vu.nl (Andy Tanenbaum) (07/11/89)
In article <5870010@hpcvca.CV.HP.COM> charles@hpcvca.CV.HP.COM (Charles Brown) writes: >Just a nit... Why does the textbook need to be restricted to 1 semester? The standard OS course is 1 semester, so the book ought to at least be usable for that. Andy Tanenbaum (ast@cs.vu.nl)
dhesi@bsu-cs.bsu.edu (Rahul Dhesi) (07/12/89)
I think Dr. Tanenbaum should consider finding a different publisher for
the revised Minix book and disks.
What Minix badly needs is a publisher that knows about software and how
to handle upgrades in an organized manner. Prentice Hall definitely is
not it.
--
Rahul Dhesi <dhesi@bsu-cs.bsu.edu>
UUCP: ...!{iuvax,pur-ee}!bsu-cs!dhesiast@cs.vu.nl (Andy Tanenbaum) (07/12/89)
In article <8176@bsu-cs.bsu.edu> dhesi@bsu-cs.bsu.edu (Rahul Dhesi) writes: >I think Dr. Tanenbaum should consider finding a different publisher for >the revised Minix book and disks. No. (1) I have a contractual obligation to P-H and would certainly be sued if I tried to breach it, and (2) although they have screwed up the software distribution, they have done ok on the book. If I were to find a software publisher, I bet they would screw up on the book. Nevertheless, P-H is definitely learning about software. They now have a full-time customer service software person, for example, and I will try to train him on MINIX next time. VISION ON THE FUTURE OF EDUCATION: I think that in the coming decade, many universities will expect/require all students to buy a computer. If you take a $2000 computer and write it off over four years, you get $500 per year. When added to private college tuition of $10K per year, it only adds 5% and the student gets to keep the computer afterwards. At state schools it will take a little longer, but not much. These machines will need software for physics labs, biology labs, library access etc. It is my bet that 99% of it will be written by free lancers, mostly professors, the same way it is with books. There are very few corporations that churn out textbooks as their main product. Thus I expect college software to go essentially the same route as college textbooks. In this case, it makes sense that the distribution is handled by book publishers. Whether they are selling a book or a box of diskettes hardly matters--the manufacturing is contracted out to third parties in both cases. It is just a 7" x 10" x 2" lump with an ISBN number in both cases. Prentice-Hall is far and away the biggest (and as far as I am concerned) the best computer science textbook publisher, and I expect them to be the same in educational software as well. I know plenty of other authors, and have heard about other people's experiences, and while P-H is certainly not perfect, I am by-and-large pretty satisfied. If any of you are thinking of writing textbooks or educational software, they are definitely worth considering. I guess it all boils down to what MINIX is. I still see it as primarily an educational system. I can't imagine any Fortune 500 company choosing it over UNIX. Correct me if I am wrong, but I think it falls under the heading Educational/Instructional rather than Operating systems (see page 336 of the July Byte Magazine). Andy Tanenbaum (ast@cs.vu.nl)
des@berlioz (Desmond Young) (07/12/89)
In article <1989Jul8.171336.27399@sun.soe.clarkson.edu>, jk0@image.soe.clarkson.edu (Jason Coughlin) writes: > From article <2835@ast.cs.vu.nl>, by ast@cs.vu.nl (Andy Tanenbaum): > > ,and perhaps > > eliminating the cross reference listing. > > ... the cross reference listing is VERY handy. I > used it ALL of the time ... Put the XREF on disk, the buyer can print it out. I used it a lot too, but it was a nuisance turning to the back of the book constantly. Putting it on disk saves book pages, and gives a physically separate reference. Des. -- Reply: des@logic.nsc.com {decwrl,hplabs,sun,pyramid,amdahl}!nsc.com!logic!des Des Young, M/S D3635 National Semiconductor, Santa Clara The light at the end of the tunnel was only a burglar's torch - J.Buffet.
rheffel@cs2.wsu.edu (07/12/89)
I suggest that AST publish the operating system book and the code listing separately. By code, I mean that Appendicies B-E should comprise the separate code listing. Perhaps even Appendix A should be included here too. The advantage arising from this separation is that the book will not contain older versions of the code (thereby rendering it slightly obsolete) and will allow operating system classes to reference current code. Furthermore it will allow MINIX devotees to keep up to date by permitting them to purchase the code listing separately. I suggest that the code listing be published in an IBM 3-ring binder size format Let us take a look at an example of what I mean. If AST's operating system book did not use line numbers but allowed one level of indirection, his book would not go out of date. For example on page 243 he references line 5965. In another version of MINIX (ie later than 1.1), that line number probably does not reference the same C statement. This problem could be solved properly by inserting the phrase "ref-listing x" in the book, and on the first page of the IBM 3-ring binder have the following: ref-listing x = line 5965. Such a method of indexing so commonly found in computer science would preclude the necessity of constantly modifying the book to keep it up to date but would still permit the OS book to correctly correspond to the lastest version of the code. Prentice Hall could bundle the 2 items (the OS book and the IBM style code listing) to sell to student book stores, but could also sell the IBM style code listing optionally with the upgrade diskettes. I hope this method of distribution will be considered seriously because MINIX has a larger following than AST presumes. I agree with the MINIX fan who suggested that the XREF program be included with the next version of MINIX.
gert@targon.UUCP (Gert Kanis) (07/13/89)
In article <563@prles2.UUCP> meulenbr@cstw01.prl.philips.nl (Frans Meulenbroeks) writes: >[lots of text discarded] >In article <2835@ast.cs.vu.nl> ast@cs.vu.nl (Andy Tanenbaum) writes: >[...] >>Anyway, I have decided to use Bruce's code as the basis for V2.0 (subject to >>Prentice-Hall not vetoing a 900 page book). >>[...] >>When all that is done, we will have a protected mode system that is based >>on POSIX and reasonably fast for small machines. MINIX seems to be >>evolving in the direction of something more than an academic toy ... >> Sounds good, I'll bet all those with 80x86's will be most happy about this decission. But what's the future of those other versions? >>An additional note. Amiga MINIX is actually up and running. It is now >>being tested. Furthermore, MINIX 1.3 is being ported to the SPARC. >> >>Getting all these versions back together,and then POSIX-based is going to be >>a job on the order of cleaning out the Augean Stables, and I don't have a >>Hercules card. >>Andy Tanenbaum (ast@cs.vu.nl) > >Ok. That amounts to 5 different versions: >for 8088, 286/protected, ST, Amiga, SPARC. > >As far as getting these together: >This can really become a pain. >However, I would support such an integration very much. > >Frans Meulenbroeks (meulenbr@cst.prl.philips.nl) > ( or try: ...!mcvax!phigate!prle!cst!meulenbr) So we see a clear path concerning the devellopment of MINIX in the next few years. One of the goals was (apart from the teaching aspects) portability. Looking at already 5 versions one cannot deny this has succeeded. But will we have 5 different versions soon? I really hope not, that would depart from the original idea. OK , I know that everybody can hack around to make his own version, what I mean is the reference, the thing you can make your diffs to, the version of The Book. (or Books :-) ) SOME QUESTIONS I'D LIKE TO SHARE. 1)At this moment we have MINIX ST 1.1 that is said to equal MINIX PC 1.3 . So why isn't it CALLED MINIX ST 1.3 ? 2)Why is it impossible to read disks written on the other version although the original OS'es of PC and ST can do so easily? (don't know about these other versions). The (ST) documentation says: The goal is not TOS (Atari) but UNiX compatibility. What about making MINIX compatible with itself. (I know someone said this before). I hope 2.0 will again become available for many computers. There is more than this PECEE. (You might have guessed I have an Atari :=) ) Don't think I'm unhappy with MINIX but this keeps bothering me. +-----------------------------------------------------------------+ | The smoker you drink,| Gert Kanis, SWP | | the W.C. | Nixdorf Computer BV, Postbus 29 | |----------------------| 4130 EA Vianen, Netherlands. | | I do not represent | E-mail : targon!gert@nluug.nl | | anyone elses opinion.| or ..!mcvax!targon!gert | +-----------------------------------------------------------------+
charles@hpcvca.CV.HP.COM (Charles Brown) (07/14/89)
>Let us take a look at an example of what I mean. If AST's operating system >book did not use line numbers but allowed one level of indirection, his book >would not go out of date. For example on page 243 he references line 5965. >In another version of MINIX (ie later than 1.1), that line number probably >does not reference the same C statement. > >This problem could be solved properly by inserting the phrase >"ref-listing x" in the book, and on the first page of the IBM 3-ring >binder have the following: > ref-listing x = line 5965. How about inserting comments in the source which you can easily search for. So the book can refer to a point in the source with: see listing-reference-232 In the source we could find if( ... ) { ... } /* listing-reference-232 */ This way the book (and the binder) would not have to become obsolete because the line numbers changed. -- Charles Brown charles@cv.hp.com or charles%hpcvca@hplabs.hp.com or hplabs!hpcvca!charles or "Hey you!" Not representing my employer. "grep for it"
rbthomas@athos.rutgers.edu (Rick Thomas) (07/14/89)
I believe that the whole problem with P-H sending people 1.2 is that
they are not trained to look at little things like numbers in book
titles.
The way I have found to be sure fire for getting exactly what you want
is to locate a friend who has the right version and get the ISBN
number off of it. Then order by ISBN number, not by title.
For example, I have the 1.3 AT version in my hand right now and its
ISBN number is 0-13-58303-5.
Accept no substitutes!
Rick
--
Rick Thomas
uucp: {ames, cbosgd, harvard, moss, seismo}!rutgers!jove.rutgers.edu!rbthomas
arpa: rbthomas@JOVE.RUTGERS.EDU
Phone: (201) 932-4301wbeebe@rtmvax.UUCP (Bill Beebe) (07/14/89)
In article <2858@ast.cs.vu.nl> ast@cs.vu.nl (Andy Tanenbaum) writes: > [ stuff deleted ] >A far more interesting question is getting the second >release of Bruce Evans' stuff and the Atari in sync, since that will be the >base for V2.0 on the IBM line. That will indeed be interesting. The Mototola chip that comes closest to the 286/386 with built-in protection is the 68030, and the only machines I know of that have that chip are Sun's, Apollo's, NeXT's, and Apple's. Not exactly in the same price range as the Atari [grin]. As the old Chinese curse says, "May you live in interesting times."
dtynan@altos86.Altos.COM (Dermot Tynan) (07/15/89)
OK, some various unsolicited comments...
1. The size of the Minix textbook.
While I don't have a copy in front of me, it would seem to me that you
could free up a lot of space by removing the "man" pages. They should
really be online like the rest of the pages, and if anyone really wants
a hard-copy, there's always 'lp'. Failing that, P-H could have a subsidiary
market for man-pages :)
2. The cross-reference listing...
I agree with a lot of what has been said already, although I have had
problems in the past, with routines that are in the code (1.2x), but not
in the book. Of course, 'grep' knows what to do in cases like that...
I don't think that indirect references in the text buy you anything,
because there is more to an update than changed line numbers. There is
no way to get around the problem of discussing a section of code that
now does something *completely* different.
3. Merged source code.
I think it would be great if all the different Minix varieties were on
the same disk-set. I'd like to see the Atari code, even if I have a
PC. On the one hand, this is easier for P-H, because they don't have
to stock multiple distributions, but requires that the distribution be
bigger. Of course, there's always the problem of different disk formats...
One thing I would like to warn against, though, is heavy use of '#ifdef's.
If they are used more than minimally, the code becomes unreadable. One
good solution is to produce a source-tree, with directories of 'common',
'IBM-PC', 'i8086', 'Atari', 'm68k', etc. The breakdown being by machine,
cpu, and common.
4. Protected mode.
I'm glad to see that Andy has decided to use Bruces' code. I hope this
doesn't mean us unfortunates stuck with prehistoric Intel chips, won't
get left by the wayside. Does it also mean that we'll see swapping (or
at least shared-text) in the next release??
5. TCP/IP.
I am going to be posting Des Youngs' Minix port of TCP/IP (KA9Q), within
the next week. Des gets all the credit for this one, all I'm doing, is
posting it on the net, because he is out of town. He put a lot of work
into it, and it's a real nice port. However, the disk he gave me has
some original files on there, so I need to diff them against the release,
so as not to infringe on P-Hs' copyright. That could take a while...
The size, by the way, is about 800K. But that is before the 'diff'ing.
- Der
--
dtynan@altos86.Altos.COM (408) 946-6700 x4237
Dermot Tynan, Altos Computer Systems, San Jose, CA 95134
"Far and few, far and few, are the lands where the Jumblies live..."ast@cs.vu.nl (Andy Tanenbaum) (07/15/89)
In article <575@targon.UUCP> targon!gert@nluug.nl (Gert Kanis) writes: >1)At this moment we have MINIX ST 1.1 that is said to equal MINIX PC 1.3 . >So why isn't it CALLED MINIX ST 1.3 ? Johan and I felt it strange to call the first release 1.3. Maybe it isn't. Also, it is not exactly 1.3. >2)Why is it impossible to read disks written on the other version although >the original OS'es of PC and ST can do so easily? I am not really sure. Probably we slipped up. The Amiga version should be able to read MINIX-ST disks, so we have tried to avoid that mistake again. >I hope 2.0 will again become available for many computers. There is >more than this PECEE. (You might have guessed I have an Atari :=) ) It may be a long road. What I do realistically hope to do is have FS and MM be absolutely identical on all versions. That is one of the advantages of splitting them off from the kernel. After V2.0 is done for the PC, I hope we can find some ambitious volunteer to try to retrofit the ST and Amiga kernels to the V2.0 kernel, where possible. I don't know if this will work. The SPARC is yet another story, since the register window business is quite different than all the other machines. Andy Tanenbaum (ast@cs.vu.nl)
pcm@iwarpj.intel.com (Phil C. Miller) (07/17/89)
In article <3534@altos86.Altos.COM> dtynan@altos86.Altos.COM (Dermot Tynan) writes: > >OK, some various unsolicited comments... and an unsolicited response. >3. Merged source code. >I think it would be great if all the different Minix varieties were on >the same disk-set. I'd like to see the Atari code, even if I have a >PC. I personally think that this is a REALLY BAD IDEA (no offense, Der). Making the number of distribution disks larger when it's not necessary seems inherently evil. This would raise the price of the disks when few people would get the benefit. > On the one hand, this is easier for P-H, because they don't have >to stock multiple distributions, but requires that the distribution be >bigger. Of course, there's always the problem of different disk formats... Precisely. Phil Miller {...}!tektronix!ogccse!pcm
vxp6840%ritcv@cs.rit.edu (07/19/89)
In article <4658@omepd.UUCP> pcm@iwarpj.UUCP (Phil C. Miller) writes: >In article <3534@altos86.Altos.COM> dtynan@altos86.Altos.COM (Dermot Tynan) writes: >... >>3. Merged source code. >>I think it would be great if all the different Minix varieties were on >>the same disk-set. I'd like to see the Atari code, even if I have a >>PC. > > I personally think that this is a REALLY BAD IDEA (no offense, Der). > Making the number of distribution disks larger when it's not > necessary seems inherently evil. This would raise the price of > the disks when few people would get the benefit. > >... I think putting all the source on one disk set is a good idea. Minix is meant to be an OS for learning. It is very informative not only to see how Minix works in my own system, but what differences exist to make it work on other systems as well. When I bought Minix for my ST, I was disappointed to find only the ST code on the distribution disks for the above reason. -Vitas P.- -Vitas P.- ...![allegra, seismo, or rutgers]!rochester!rit!vxp6840 Rochester | vcp6840@ritvax.bitnet or vxp6840@ritcv.UUCP Institute | vcp6840@ultb.isc.rit.edu or vxp6840@cs.rit.edu of Technology | SOMETHING HAS TO WORK AROUND HERE!!!!!!!!!!!!!!!!!!!!!!
ONM64%DMSWWU1A.BITNET@cunyvm.cuny.edu (Kai Henningsen) (07/19/89)
Well, here it's again ... not a good Idea to send it to MINIX-L@DEARN.BITNET, INFO-MINIX@UDEL.EDU is the place if I wish to reach the rest of the world ... >From: Andy Tanenbaum <ast@CS.VU.NL> >VISION ON THE FUTURE OF EDUCATION: >I think that in the coming decade, many universities will expect/require >all students to buy a computer. If you take a $2000 computer and write it >off over four years, you get $500 per year. When added to private college >tuition of $10K per year, it only adds 5% and the student gets to keep the >computer afterwards. At state schools it will take a little longer, but not >much. Whow, that's numbers! I'm really glad I live in Germany ... that's about DM 50,- (ca. $25) per Semester ... $10K??????? ~
craig_dewick%713.600@fidogate.fido.oz (Craig Dewick) (07/31/89)
Original to: ast@cs_vu_nl Hello Andy. It is pleasing to see all this support for Minix. I own a kit 68000 system for which a port of minix 1.2 has been recently completed. The final code looks nothing like your original listing in the 'red book', however it works and performs well. It has been over a year in development, mainly because the person who did the port has not been able to spend all his time on it. There are still many small problems to be ironed out, but very soon we should have a very stable, efficient version of Minix running on our machines. I have never used Minix on a PC so I cannot say how it compares performance-wise, but I am looking forward to a new book. I have read the current edition cover to cover (except the source code) and found it very informative. I have used Unix on the university's Amdahl mainframe and have never really understood many of the things the experienced Unix users take for granted. Your book helped me understand some of these things. I will be eagerly awaiting the new edition. C ya later.... Craig. --- TMail v1.11b * Origin: Prophet TBBS - First TBBS in Australia (3:713/600)
cwr@pnet01.cts.com (Will Rose) (02/15/90)
It's been quite instructional to watch the growth of Minix over the last few years, especially to one interested in business and social systems, as well as computing. It seems that PH has now realised that it is secure in its niche, and can charge accordingly. (No-one with the skill will ever have the time to build another Minix; if they did, it would have great difficulty getting established against the current version.) Certainly $120 for a program with no development cost (subsumed into the original textbook cost), no support cost (apart from printing ast's email address), and no maintenance cost (partly free from the net, and partly subsumed into the cost of ast's next edition) is good business. My guess is that 1.6 and later will allow no copying, and PH will try to move Minix as far upscale as the market will bear. The fun will be to see how far they get, without falling flat on their faces or pulling in the competition. In fact, the figures may be better than that. I bought the slip-case with Minix 1.2, (I already had the textbook) and the slipcase stuff was fairly useless. Most students probably buy the textbook already - it's a very good one; they will all still need the book, and so adding other material to disk sales shouldn't cannabalise text book sales at all. However, at the end of the day student's using Minix will be paying $120-130 for the program, $40-$50 for the textbook, and $20-$30 for a book on using Unix (eg. McGilton and Morgan). A lot of money for one course, or even two; I wonder what will happen next? Astonishing good luck, Sirs, and remember us as, the Filthistan Trio Will Rose ========================================================================== | Joseph Schumpeter lives! | cwr@pnet10.cts.com | ========================================================================== UUCP: {nosc ucsd hplabs!hp-sdd}!crash!pnet01!cwr ARPA: crash!pnet01!cwr@nosc.mil INET: cwr@pnet01.cts.com
jk0@image.soe.clarkson.edu (Jason Coughlin) (02/15/90)
Someone said: > Unfair to shell out $115 for one semester? That's not that bad, in comparison >to some $70 chemistry text books I have used. A hell of alot more usefull >once the course is finished! This is exactly what I'm trying to avoid -- paying $115 for a program that won't be used the minute the final is over. (See below for why the program isn't used later.) Someone else said: >EE/CS 338 at UIUC: Computer Networks (A undergrad/grad course): >There are two required texts (one of which is by Andy Tanenbaum), which the >bookstore has ONLY NEW copies of: the Tanenbaum book is $55, and the other is >$59. I spent $250 for books last semester. So what is your point? You didn't *want* to pay $250 did you? (Answer 'No'). Then why would you want to spend $115 on Minix? My point is to prevent this kind of thing. From article <1500@crash.cts.com>, by cwr@pnet01.cts.com (Will Rose): > However, at the end of the day student's using Minix will be paying > $120-130 for the program, $40-$50 for the textbook, and $20-$30 for > a book on using Unix (eg. McGilton and Morgan). A lot of money for > one course, or even two; I wonder what will happen next? Don't forget the $250 (and that's an underestimate) for a hard-disk, Minix 1.5, 1.6, 2.0 all require a hard-disk. Now come off of it: $40 + $115 + $250 = much too much for an *undergrad* course with a program that is a "learning" OS. Undergrads just don't spend enough time digging through code to make $115 worth it. > Certainly $120 for a program with no development cost (subsumed into the > original textbook cost), no support cost (apart from printing ast's email > address), and no maintenance cost (partly free from the net, and partly > subsumed into the cost of ast's next edition) is good business. My guess > is that 1.6 and later will allow no copying, and PH will try to move Minix > as far upscale as the market will bear. The fun will be to see how far > they get, without falling flat on their faces or pulling in the competition. > I concurr. Remember folks, Minix is a LEARNING OS -- developed to *TEACH* the ideas of OS. The program is not a "real" enough OS to replace MS-DOS, OS/2, or AIX so when the course is over, Minix most likely goes back in the disk box. If Minix were totally capable of replacing MS-DOS, OS/2, or AIX (on the student's machine) then $115 is a bargain that couldn't be passed up. BTW, my opinions are based on $115 price for Minix. They are NOT a criticism of Andy's work or of Andy in general. Personally, I like the beauty of Minix. -- Jason Coughlin ( jk0@sun.soe.clarkson.edu , jk0@clutx ) "Every jumbled pile of person has a thinking part that wonders what the part that isn't thinking isn't thinking of." - They Might Be Giants
jds@mimsy.umd.edu (James da Silva) (02/15/90)
In article <1500@crash.cts.com> cwr@pnet01.cts.com (Will Rose) writes: >I bought the slip-case with Minix 1.2, (I already had the textbook) and >the slipcase stuff was fairly useless. I'll go farther: The slipcase version is COMPLETELY useless to the user; its sole reason for existence is to provide the bookstores with the proper margin. It isn't even the right size! Down with the Slipcase version! Up with The Book + The Disks! Jaime ........................................................................... : domain: jds@cs.umd.edu James da Silva : path: uunet!mimsy!jds Systems Design & Analysis Group
peter@ficc.uu.net (Peter da Silva) (02/16/90)
In article <1990Feb15.153248.6537@sun.soe.clarkson.edu> jk0@image.soe.clarkson.edu (Jason Coughlin) writes: > I concurr. Remember folks, Minix is a LEARNING OS -- developed > to *TEACH* the ideas of OS. The program is not a "real" enough OS to > replace MS-DOS, OS/2, or AIX so when the course is over, Minix most > likely goes back in the disk box. If Minix were totally capable of > replacing MS-DOS, OS/2, or AIX (on the student's machine) then $115 is a > bargain that couldn't be passed up. [AIX? Why bring up AIX? How about System V/386?] That depends on what the student wants to do. The majority of computer science or engineering types I know who have PCs use their computers for light editing, programming/hacking, modemming, and running BBSes. For this activity Minix is far superior to DOS or OS/2. And for these people, Minix is indeed a bargain. -- _--_|\ Peter da Silva. +1 713 274 5180. <peter@ficc.uu.net>. / \ \_.--._/ Xenix Support -- it's not just a job, it's an adventure! v "Have you hugged your wolf today?" `-_-'
sung@cs.washington.edu (Sung Kwon Chung) (02/16/90)
[Discussions regarding the (expected) higher costs (text, software, hard-disk, ... ) of using new versions of Mnix in OS classes] This is not a new idea, but how about using earier version (1.2 or even 1.1) in classes? After the class, students with further interest in Minix can purchase the newest version. (I am almost assuming the earier versions of Minix is freely copiable to students. But I maybe wrong.)
ast@cs.vu.nl (Andy Tanenbaum) (02/16/90)
In article <1500@crash.cts.com> cwr@pnet01.cts.com (Will Rose) writes: >Certainly $120 for a program with no development cost (subsumed into the >original textbook cost), no support cost (apart from printing ast's email >address), and no maintenance cost (partly free from the net, and partly >subsumed into the cost of ast's next edition) is good business. My guess >is that 1.6 and later will allow no copying A couple of messages back someone noted that he had just bought a chemistry book for $70. I suspect that if that publisher had to throw in 17 diskettes as well, they wouldn't come out that much lower than $120. Furthermore, I bet they wouldn't permit their author to distribute free upgrades to the disks indefinitely on the net. Version 2.0 and beyond will have the same license as the current one: you may may 2 copies, so three people can share the $120. I think compared to what Microsoft, Lotus, and most other companies charge, this is not outrageous. > I bought the slip-case >with Minix 1.2, (I already had the textbook) Why did you buy the slipcase version when the yellow box with the disks was available for much less? The slipcase version is aimed at the professional market. When MINIX was first announced, a couple of hundred people a day came running into the Computer Literacy Bookstore in Silicon Valley wanting to buy the software. The owner called P-H about twice a day bugging them for a version he could sell. That's when P-H made the slipcase version. There are no plans at all to change the current approach, except perhaps to shift the documentation from the book to the manual that comes with the disks to reduce the size of the book. I will keep posting upgrades to the net so anyone who wants can have them for free. Andy Tanenbaum (ast@cs.vu.nl)
anderson@vms.macc.wisc.edu (Jess Anderson) (02/17/90)
In article <5443@star.cs.vu.nl>, ast@cs.vu.nl (Andy Tanenbaum) writes... ]A couple of messages back someone noted that he had just bought a chemistry ]book for $70. I suspect that if that publisher had to throw in 17 diskettes ]as well, they wouldn't come out that much lower than $120. Furthermore, I bet ]they wouldn't permit their author to distribute free upgrades to the disks ]indefinitely on the net. Version 2.0 and beyond will have the same license ]as the current one: you may may 2 copies, so three people can share the $120. ]I think compared to what Microsoft, Lotus, and most other companies charge, ]this is not outrageous. Case in point: yesterday my upgrade from the standard edition of Microsoft Word to Word for Windows arrived. My employer (who escaped taxes by virtue of being a government) shelled out $150 plus $5.50 shipping. This brings 7 or 8 floppies and a set of documentation. To my utter stupefaction, this $155 investment included a card I could send in -- with (you guessed it) *another* $20 so I can have the technical reference manual (with such useful things as how you write macros, etc.). It's reported to be a superior product (I can say it's no great speed demon, despite my having a 33 MHz 386 machine with 8 megs, a 17ms hard disk, a 256K ram cache and a 1.5M disk cache). But compared to Minix? Let's be realistic, a working operating system--contantly improving and expanding at an amazing rate--plus a first-rate textbook by an author who is himself a simple email message away, who proves his star status by never acting like a star, plus the most liberal licensing conditions in the entire industry, all this for less than 25% of the cost ($495) of Word for Windows. Let's face it, there are people who will complain about the price if it's free or even if you pay *them*, and students probably have some legitimate concerns, given the debt some of them routinely assume as part of getting the degree, but despite that, Minix is still the deal of the decade. ==Jess Anderson===Academic Computing Center=====Univ. Wisconsin-Madison===== | Work: Rm. 2160, 1210 West Dayton St., Madison WI 53706, Ph. 608/263-6988 | | Home: 2838 Stevens St., 53705, 608/238-4833 Bitnet: anderson@wiscmacc | ==Internet: anderson@macc.wisc.edu====UUCP:{}!uwvax!macc.wisc.edu!anderson==
jk0@image.soe.clarkson.edu (Jason Coughlin) (02/18/90)
From article <5443@star.cs.vu.nl>, by ast@cs.vu.nl (Andy Tanenbaum): > A couple of messages back someone noted that he had just bought a chemistry > book for $70. I suspect that if that publisher had to throw in 17 diskettes > as well, they wouldn't come out that much lower than $120. Furthermore, I bet > they wouldn't permit their author to distribute free upgrades to the disks > indefinitely on the net. Version 2.0 and beyond will have the same license > as the current one: you may may 2 copies, so three people can share the $120. > I think compared to what Microsoft, Lotus, and most other companies charge, > this is not outrageous. Since I started this thread, I would like to finish it before we get into a war. My whole point was that I thought it was unfair to make a student pay ($120 + $40) for one undergraduate level OS course. I didn't say that $120 was unreasonable for MINIX *period* -- I said it was unreasonable to charge for an *undergrad* OS course. Andy said that educational institutions are allowed to make copies for their students taking the class. THEREFORE, the war is mute. -- Jason Coughlin ( jk0@sun.soe.clarkson.edu , jk0@clutx ) "Every jumbled pile of person has a thinking part that wonders what the part that isn't thinking isn't thinking of." - They Might Be Giants