ast@cs.vu.nl (Andy Tanenbaum) (05/28/91)
There has been a lot of discussion about "Advanced MINIX" and other related topics lately. A great deal of it is based on hearsay and some people's vivid imaginations. Although I have said most of this before, let me say it again to try to set the record straight. PH is not trying to steal everybody's software. They are only marginally interested in the whole MINIX thing, and if it weren't for the fact that I nag them once in a while, they would just let the whole thing die off peacefully. The image of them as some rapacious corporation feeding off the hard work of all those dedicated hackers in netland is absurd. If I were to ask them: "Should I freeze MINIX at 1.6.16 and never take another line of code from anyone or change anything myself?" the answer would certainly be yes. The current listing is at least 500 pages, which means that the second edition of the book will be > 1000 pages. Neither they nor I like this, nor do we like having 2 500-page books, or 3 333-page books, or 4 250-page books, or 100 10-page pamphlets. To the extent that I have time and think it is fun, I am working slowly towards IEEE P1003.1, mostly because I believe in standards. Americans living in Europe are probably more sensitive to standards than anybody else. People in the U.S. just assume that the whole world uses inches; it never even occurs to them that the U.S. way in many areas is weird. Europeans just shrug and assume that it is natural that every time you drive 1000 km the shape of the electrical outlets change, so a device that works at (x, y) can't be plugged in at (x + 1000, y). Nevertheless, given that PH has promised me to try to make some effort to take MINIX slightly seriously, they don't want to have to compete with Fred van Kempen selling what is largely their own product. (A year or so ago, Fred implored PH to let him become the official dealer for the Benelux countries, which they agreed to, but when he sold few and paid for none, they dropped him in favor of a local store.) I guess this made him unhappy. Is revenge a legitimate motivation for writing software? Who knows. In any event, neither they nor I mind him or anyone else modifying MINIX and posting cdifs or putting them on archives. What they do mind is his making a self-contained version that people can get instead of theirs. An add-on is ok, but a replacement is not. Seems pretty clear to me. Fred says he sells it only to people who can prove they have real MINIX, but that is awfully hard for them to verify. Suppose something buys MINIX, makes 20 copies of the receipt, and gives them to his friends. Suppose somebody makes 20 color photocopies of the boot disk label and glues them to blank disks. PH does not want to be dependent on Fred's doing a really thorough job of checking out people--after all he has no real interest in doing so. Suppose somebody took the Microsoft C compiler or Lotus 123 or Word Perfect, disassembled it, modified the assembly code, and began selling it or even giving it away, but only to people to could prove they had legally bought the original. How do you think the maker would react? I suspect in an unfriendly way. I also suspect that they would not approve of cdiffs either, which PH doesn't mind. Thus, I think PH allows a lot more than any other company selling any kind of software. I think it is also worth pointing out that I have been providing free upgrades via the net since the beginning, and intend to continue doing so. Do Microsoft, Lotus, Word Perfect, etc give free upgrades? If you want to think of this as a kind of "repayment" to the net for contributed software, fine, but if you compare PH to other companies selling software, I think they are a lot less commercial than most. For a more commercial approach, look at comp.os.coherent. People also provide programs, ideas, feedback etc., there to the Mark Williams Company, but MWC does not provide source code or free upgrades. I don't see why Fred can't put his cdiffs on the archives and not insist on the right to distribute self-contained disks, including all of my copyrighted code. It may be that the cdiffs are big, but patch doesn't care, and I haven't heard any archive maintainers complain about all the space this would take. The actual work of combining 1.5 with his patches could be in a shell script. It would allow anybody who wants to upgrade to do so, and be legal. What's wrong with that? Why can't he do it the legal way? It's hardly more work. Fred has said that if he doesn't get the right to distribute PH's copyrighted code, he will write his own MINIX clone. If the clone is entirely his, and does not steal from me or other people, that is his right. He is trying to enlist supporters to work on this project. As a point of information, it would be interesting to know what machines this project is targeted for. If it is 386 machines (e.g., because nobody has a good PD compiler that runs on the 8088), people should be aware that in about 3-4 months, Berkeley will be releasing 4.4 BSD for the 386, in source form and totally free of AT&T code. Thus people who want an AT&T-free UNIX for the 386 with source code will be able to get it. In that light, putting in a lot of effort to rewrite MINIX to make it Berkeley UNIX seems to be a bit of a wasted effort. Also, this new system will probably be directly competing against a variety of other high-end systems such as 4.4, Mach, GNU, Chorus, Amoeba, etc. It's going to be a tough sell for Fred. My niche will continue to be educational users and people who want a system that is small enough to understand. Given the Berkeley tradition, 4.4 is not likely to be real small. Maybe I shouldn't have said all this, but it is beginning to irritate me that a lot of people are accusing PH of being so predatory, when in fact they are hardly commercial at all. It's taken me two years of discussion, but they are finally going to run a second ad in Byte in June or July. The ad itself is a real turkey.
tsarna@polar.bowdoin.edu (Tyler Sarna) (05/29/91)
Andy Tanenbaum wrote: > having 2 500-page books, or 3 333-page books, or 4 250-page books, or 100 > 10-page pamphlets. 21,000 playing cards? > IEEE P1003.1, mostly because I believe in standards. Americans living in > Europe are probably more sensitive to standards than anybody else. People > in the U.S. just assume that the whole world uses inches; it never even > occurs to them that the U.S. way in many areas is weird. Europeans just shrug Well, that's a pretty sweeping statement, but I'll agree in general. Personally, I like the SI much better than the English system, and I certainly beleive that the US is weird in more than a few ways. > and assume that it is natural that every time you drive 1000 km the shape of > the electrical outlets change, so a device that works at (x, y) can't be > plugged in at (x + 1000, y). Boy, you must have very, very straight roads in the NL if that math works :-) > for them to verify. Suppose something buys MINIX, makes 20 copies of the > receipt, and gives them to his friends. Suppose somebody makes 20 color > photocopies of the boot disk label and glues them to blank disks. PH does I don't see that the logic holds -- nothing prevents the person from applying the patch kit to his copy of MINIX, and THEN giving it to 20 friends. I do agree, however, that it would be much easier for FvK to just distribute the cdiffs than to try to be some sort of martyr. One wonders why he doesn't. > [A new un-Minix not a bright idea] Yes, I beleive I made this point in an earlier post. The freenix/cheapnix market is rather crowded and will get even more crowded soon. Yet another new system serves no purpose, other than the joy of creation. -- Tyler "Ty" Sarna tsarna@polar.bowdoin.edu "Death therapy, Bob. It's a guaranteed cure."
klamer@mi.eltn.utwente.nl (Klamer Schutte) (05/30/91)
In <54878@nigel.ee.udel.edu> tsarna@polar.bowdoin.edu (Tyler Sarna) writes:
%>Andy Tanenbaum wrote:
%>> and assume that it is natural that every time you drive 1000 km the shape of
%>> the electrical outlets change, so a device that works at (x, y) can't be
%>> plugged in at (x + 1000, y).
%>Boy, you must have very, very straight roads in the NL if that
%>math works :-)
Euclidean math won't help you. 1000 km and the Netherlands -- any map will
show what a funny mistake this is ;-)
Klamer
--
Klamer Schutte Tel: +31-53-892786 Fax: +31-53-340045
Faculty of electrical engineering -- University of Twente, The Netherlands
preferred: klamer@mi.eltn.utwente.nl SMTP: klamer@utelmi01.el.utwente.nl