Slashdot Mirror
Do You Know UNIX Secrets?
ESR writes "You can help stop the SCO attack on IBM and the Linux community.
I'm looking for ways to prove that Unix trade secrets have been legally
nullified.
I want to know if you have ever had read access to proprietary Unix
source code (not just binaries and documentation) under circumstances
where either no non-disclosure agreement was required or whatever
non-disclosure agreement you had was not enforced. To help out, see my No Secrets page."
This is the book which shows the roots of Unix: Lion's Commentary on UNIX 6th Ed. with source code
When Sun gave away Solaris source code way back...did that maybe justify the cause?
-psy
It the difference between showing a signed petition with thousands of signatures, vs citing a poll that X% of the constituents is for or against somethng.
To run for office you need signed patitions not just a few pollsters saying you are popular.
Help fight continental drift.
If the claim is that the system needs to be stricken of all SCO *owned* code. Isn't the first argument: What code do you really own, what happened in the AT&T Case (open the court files) show us what needs to be changed BEFORE a lawsuit ever took place.
The technical community is VERY good at solving these sorts of problems. Even if we are talking 6-18 months time to rewrite all the sources they are claiming are in violation. That's much shorter than this court case with IBM is going to be.
Besides, even if they are not *really* in violation. And there are claims for both sides, if they were re-written, SCO loses their lawyer money, and we can go back to ragging on M$-Wintel-DMCA and the US court system.
During the times I was involved with actual Unix source code (mid 1970's to mid 1980's) for three different employers (UCLA, SDC, Tektronix) there was nothing similar to a nda. The school or company just agreed to not give source code access to non-"employees", and "employees" agreed not to give access to others who hadn't agreed to the Bell license.
Within those terms, there was a lot of access. At the yearly conferences (which later became USENIX) there was a typically conference distribution tape. That tape was a mixture of "new" things, and modifications to the Unix kernel or Unix commands. To assure that everybody was "licensed", when you first became a member you submitted a copy of the signature page of the license.
During that time we went from V6 Unix, PWB Unix, V7 Unix, to 2BSD (pdp11) and 4BSD (vax).
Sharing went both ways of course. A number of changes/new cmds from other groups became part of the Official Bell release.
That sharing was a factor in the settlement of the USL vs UCBerkeley lawsuit, that ended in the free availability of the 386bsd work.
I was at Bell Labs for almost three years in the early 1980's, moving over to AT&T Information Systems after the court ordered breakup of AT&T in 1984 or so.
They were pretty laid back then; I may have signed an NDA but I certainly don't recall it. I do recall the usual W4 and Insurance BS but an NDA doesn't stick out.
And yes, I had almost full access to the source tree. IIRC, only some arcane kernel stuff wasn't available, being crafted in assembly. But given the corporate culture I have no doubt it was somehow accessable, but because it was processor / architecture specific, I never bothered looking for it. Plenty of stuff to look at and learn from at higher levels.
Source code was available to any member of technical staff and since it was my second job out of Uni I had a ball. I even dl'ed some source to my Osborne I so I could read it at my lesiure.
In fact I didn't realise how special it was at the time to have access to Unix source code until maybe five years later when I'd moved over to Wall Street.
The Street was ramping up sharply on tech in those days, and Unix (think Sun, NeXt and SGI workstations) was the only game in town since PCs were still pretty underpowered.
I remember someone asking me a question, and I told him to "grep for it". He looked at me cryptically, and then it hit me.
No way to grep Dude - they's binary distributions.
A message from our sponsor
Problem #1: Just because it wasn't enforced at one time, doesn't mean the NDA is null and void
No, it doesn't. But if you can prove that the Unix-company in charge of those NDAs actually did not protect their trade secrets, that is very nice. If they released it into the public by neglect of their own NDA, they are at fault. Now if the signer of that NDA did anything wrong (regardless of sloppy enforcement or not), that would be another matter...
Problem #2: Disproving "trade secret" status is pointless(not to mention, unlikely to happen). It's still copyright SOMEONE ELSE, and YOU CAN'T USE IT unless they let you!
No, but you can write your own implementation of anything copyrighted (it's not a patent), something you could not legally do with a trade secret you have knowledge of. Now if SCO is claiming that IBM went copy-paste, you would be right. But last I heard they were throwing around FUD like "stealing ideas from proprietary SCO code and incorporating into Linux". To gather all information that was, is and has been publicly available, you show that there was in fact not a trade secret and so, nothing could have been stolen.
Problem #3: It's been said time+time again, there is no proprietary code that belongs to SCO in the Linux kernel.
So you say. SCO says otherwise. Which is why we're going to court, isn't it? Where the hell do you get the divine knowledge to know that noone anywhere ever fell for the temptation to copy-paste a little? Oh, right you've been listening to 10,000 posts to slashdot on how that *can't* be the case. Nevermind...
By the way, why are people wasting time "helping" IBM? (...) and with legal "help" like this, who needs SCO? Dispute the claims intelligently, boycott SCO, whatever- just leave the legal arguments to the lawyers, or you just might end up helping SCO...and IBM won't be the only loser if that happens.
SCO has threatened to sue Everybody(tm) right down to your favorite Linux distro, and if you're a Linux user, you too. If you want to pretend IBM is the only one that could get hurt here, think again. The SCO execs and lawyers are going "Damn. They're calling our bluff. Again. Better send out a troll to slashdot to keep them from counting the cards."
Kjella
Live today, because you never know what tomorrow brings
I'll begin with a question....
Is ESR your wife, shrink, attorney, or priest?
(I think that covers all the exemptions from being compelled to testify) If the answer is "no" on all accounts, don't talk to him. He can in no way guarantee that you won't be compelled to testify at trial, or that you, yourself, will be safe from prosecution. This is just doing discovery for SCO's attorneys, pointing out individuals with names of people who could have tainted the Linux source. I'm of the opinion that they have absolutely no proof of this, why assist them in finding it?
I understand the community wants to do something here. There is, however, little we can do to help, other than keep our mouths shut. Want to help? Convert an AIX or SCO Unix machine to Linux. We don't want to show support for SCO in any way.
Then let the folks who sign their correspondance "Esquire" figure this out.
Once it is known what line(s) of code are involved, then it's a matter of tracing back to who wrote it or contributed it. Testimony by the author may help. Proof of release dates may help. There are ways once the specifics are known. SCO is trying to prevent anyone from having time to research this by hiding facts until they actually have to show them during the legal proceedings.
And if in fact it turns out to be genuine SCO code that got in there, somehow, it should be possible to figure out the path it came from. The code would be removed and the kernel developers could cite where it came from. Maybe someone would get in trouble, depending. But until specifics are known, no serious action can be taken. If it's userland code, then it's a matter of that project author or team to work with.
If it's SysV init script code, at least my servers are safe because I already got rid of that :-)
now we need to go OSS in diesel cars
BUT IN 1994 Novell who then fully owned UnixWare...From the Usenet Archives
Did Novell, who at that time owned the Unixware source, put some of the code into GPL'ed Linux to remain compatable with UNIX binaries?Time to dig up those old copies of Byte and PC Weekly.
It makes the Copyright statement from the terminfo file that ESR maintains seem oddly prescient. The statement is accessible on my RH8 system with (2nd line, last paragraph).
What's next? I predict Microsoft will sue Sun for Copyright infringement of
Actually, I'm sure his lawyer-type wife (as well as any other lawyer or person versed in trade secret law) can tell you why he's doing it. You see, there's this little precedent set in regards to trade secrets. If they've failed to enforce their NDA before, then they might not be able to make a claim against someone else who didn't uphold it.
I learned this when I left an employer and they realized I had worked there for ohhhhhh... about 3 years without signing their non-compete (duh!). The only thing they could sick a lawyer on me for was to protect their precious "trade secrets".
Their trade secrets were the following:
1) Their "best practices" methods of engineering, installing, and maintaining networks. Luckily, TCP/IP networking is pretty much common knowledge. The "best practices" are easily attainable from the Cisco and Microsoft websites (I helped them compile their best practices from those two websites). No way they could be considered a trade secret. This is analogous to the SCO fraudulent case as follows: They claim that the Linux OS couldn't have advanced without the inside knowledge attainable only from the UNIX source code. Unfortunately for SCO, if the "Enterprise kernel features HOWTO" is out there somewhere, SCO is screwed on this point.
2) Their special configuration and wiring of a modem to be used for a special out-of-band management device. Unfortunately, they had released the information to one of our suppliers. The supplier, in turn, posted it on their website. Doh! In the trade secret world, if one person can take a bit of information public, you can't make another person _not_ take it public. For example, BSD code is public, so they can't attack IBM for releasing a version of the same code. They tried to and LOST the case against BSD years ago. No enforcement against BSD == no enforcement against IBM.
3) Their last "trade secret" was supposedly their customer list. This one stumped me at first, because I (obviously) had knowledge about their customers. I knew who they were, their administrator/root passwords, their contact information, etc. etc. I even had a job offer or two from them. To use this information to my financial advantage _must_ be considered inappropriate use of a trade secret, right? WRONG! Their customer list is printed right on the back of one of their marketing brochures ("Look who we've done work for!"). Company name in hand, the phone book/Internet provided me with all of the contact information I wanted. As far as the customers' network layouts and passwords, those are proprietary information of the customers, not my former employers. Granted, it was a partial customer list and quite old. That didn't matter to me. The information I needed to get started was there. That was two and a half years ago. I've been stealing business from them ever since (not hard since they no longer provide the services I offered to their customers). This is much like SCO's earlier release of an old version of the UNIX source code. That earlier release could be enough to fry their case. Even though they are accusing IBM of raping their more recent versions, if the code in question is anywhere _near_ present in the old version, SCO is screwed.
Personally, I think SCO's case is going to fall... hard. If they had a real case, they would have shown the code long ago. Not just shown the code, but PROVED that they had it first and not the other way around. After all, Linux had many enterprise features before SCO's various *NIX flavors. Perhaps we should be sueing them for violation of the GPL?
-- S
-- Stu
/. ID under 2,000. I feel old now.