geoff@callan.UUCP (Geoff Kuenning) (03/27/84)
BTL UNIX is not copyrighted, it is protected as a trade secret. This means that copyright notices are not only not required, they could be interpreted as voiding the trade secrecy. If you read the standard Bell contract for UNIX (if you don't have a copy, it can be found in "The Software Legal Book" by Paul Hoffman), you will discover that it clearly states that ALL source code distributed by Bell Labs as a part of UNIX is a trade secret of Bell Labs. Further, there is a very long list of EVERY program and file that is covered under the contract. Now, trade-secret protection is a very tricky think. It can be lost simply by having the object of the secret be published. For example, if I find out the secret of Coca-Cola and print it in the LA Times, Coke Corp. can prosecute me for the billions of dollars of damages I have done them. But the secrecy has been lost, and ANYONE else can use the formula without any liability whatsoever. This means to me that Bell cannot protect any shell script or other world- readable ASCII file, regardless of what their contract says. (Note that this includes /usr/dict/words). They also cannot protect anything against programs such as "strings", or for that matter against disassembly. They CAN prosecute you for posting "dd.c" to net.sources--but, as I see it, once you have done so, they cannot prosecute anyone else for making use of that source file. There are two caveats to this last statement: first, don't get the bright idea of posting /usr/src to the net. In that kind of case, since everyone on the net is *WELL* aware that the sources are actually a trade secret, I could easily see a court deciding that anyone who made use of those sources was civilly liable for damages. Second, remember that when you post any source to the net illegally, your institution as well as yourself is liable for damages (because they are the holders of the license and are responsible for making sure you don't misuse sources). So if you want to bankrupt your employer or your university, just post /usr/src to the net and make sure Bell notices... Final caveat: I am not a lawyer or legally trained. Believe the foregoing at your own risk! Geoff Kuenning Callan Data Systems ...!ihnp4!sdcrdcf!trwrb!wlbr!callan!geoff
seifert@ihuxl.UUCP (D.A. Seifert) (03/30/84)
> Now, trade-secret protection is a very tricky think. It can be lost > simply by having the object of the secret be published. For example, > if I find out the secret of Coca-Cola and print it in the LA Times, > Coke Corp. can prosecute me for the billions of dollars of damages > I have done them. But the secrecy has been lost, and ANYONE else > can use the formula without any liability whatsoever. Depends on how you find out what the secret of Coca-Cola is. There are only three people in the world who know this secret. They thus take precautions such as not flying on the same plane together (in case it crashes). If you get the secret from Coca-cola, (ha!) sign a non-disclosure agreement, and then publish the info, yes Coca-cola can sue you. If however, you buy a bottle of Coke(tm) at your friendly neighborhood vending machine (no document signed!), and "reverse engineer" it by chemical analysis, there's nothing they can do about it. Their secret depends on you not being able to reverse engineer their product. That's what *secret* means. Surely with modern technology it would be possible to reverse engineer Coke. Appariently(sp) noone thinks it's worth the trouble, they'd rather market a slightly different taste and hope the the public likes it. UNIX(tm) is a little different. AT&T will not sell it to you unless you sign the document saying you promise not to tell. Thus if you have a legal copy of UNIX to reverse engineer, you have promised not to tell. (If you have the source, you wouldn't have to reverse engineer it...) Hmmm... what happens when UNIX becomes so popular that virtually all computer people have (legal) access to it. Not much of a "secret" anymore, eh? as if you didn't know, UNIX is a trademark of AT&T-BL, and Coke is a trademark of the Coca-Cola Company. -- _____ /_____\ I taught Walter Mitty everything he knows! /_______\ Snoopy |___| ____|___|_____ ihnp4!ihuxl!seifert
geoff@callan.UUCP (Geoff Kuenning) (04/07/84)
> Surely with modern technology it would be possible to reverse > engineer Coke. Apparently no one thinks it's worth the trouble, > they'd rather market a slightly different taste and hope the > the public likes it. -- D.A. Seifert Actually, I have heard that modern technology still isn't good enough. You can find out what the ingredients are, but you cannot figure out what order the process steps were done in, and apparently this makes a difference in the taste. (I only know what I am told). > UNIX(tm) is a little different. AT&T will not sell it to you > unless you sign the document saying you promise not to tell. > Thus if you have a legal copy of UNIX to reverse engineer, > you have promised not to tell. (If you have the source, you > wouldn't have to reverse engineer it...) Actually, this is no longer true. It is very easy to get a UNIX binary without ever signing a license. It is also possible to dial up to a public-access UNIX system (e.g., any university) and disassemble /bin and possibly the kernel. Bell (okay, AT&T since Jan 1, but I still think of them as Bell) might be able to sue the operator of that system because it is possible to run a public-access UNIX with very limited read permissions on /unix and /bin, but many people don't. Bell would probably have a problem proving *which* public-access system was the source of the disassembly (so to speak). And, of course, anybody can "disassemble" a shell script. >> Does all this discussion about trade secrets and publication imply >> that the publication of the V6 kernel by (I believe) the University >> of New South Wales made V6 public domain? Could I type in the code >> and sell it without paying Bell? > No, because you had to prove you had a UNIX source license before > getting the NSW UNIX books - i.e., if you had the books, you already > 1) paid Bell and 2) agreed not to distribute the code. - Guy Harris Not me! I got the UNIX books as a gift (I certainly won't tell who) and probably illegally, since they were Xeroxed (which doesn't affect their effect on the trade-secret issue). Furthermore, I know for a fact that at least some universities have used the Lyons books in courses and allowed the students to keep them. This would also seem to make the trade secrets nonsecret, since (1) one doubts the students were required to sign nondisclosure agreements, and (2) as I understand it, trade-secret law requires that the books be turned in when the students leave the organization with the UNIX license. Furthermore, for whatever reason, many people out there have illegitimate copies of the source tapes. I have personally seen illegal copies of V7 and System III (not mine!). Most universities have student employees who, in practice, have access to system consoles and 'root' accounts; many of these people have made tapes themselves in the wee hours of the night, often in the spring of their senior year. The same applies to commercial organizations. The 'secret' has leaked in many ways, and it would be very hard for Bell to prove who was the source of the leak in the general case. Again, in any case, all that affects is who gets sued, not the validity of the trade secret. BTW, I for one am not motivated by unhappiness with Bell's pricing to void the secret. I *would* like to see decent support by an organization which gets a large fraction of its revenue from UNIX (i.e., has a financial incentive to care about us little guys). If that happens by having UNIX wind up in the public domain, great. But in general I think that the negative effects of a public-domain UNIX (mainly loss of Bell-enforced standardization) far exceed the positive effects. And I also think Bell deserves a profit on a successful product that they have spent millions developing (I am refering to the salaries of all those "research" types who spent time writing utilities instead of inventing new flavors of transistor). Geoff Kuenning Callan Data Systems ...!ihnp4!sdcrdcf!trwrb!wlbr!callan!geoff Lawyer? Me? I can't even spell "Jeff" right!