Sometimes I think there must be a curse on UNIX, like there supposedly was on the tombs of some of the pharaohs of Egypt. If you stole any of the contents of the tombs, even though you might be tempted by the possibility of easy wealth, you ended up horribly dead or worse. Or like the curse on the city of Babylon, where anyone trying to rebuild the city would be doomed to failure.
I think owning UNIX must be like that. Maybe it was supposed to stay free, and when AT&T decided to take everyone's freely donated work and "monetize" it, the curse began. Now it looks like it's on SCO.
The invoice thing didn't work out well. The license program is on a slow boat to nowhere. The IBM trial is not helping, and Red Hat is breathing down their neck too. And now they have announced they want to collect money because of some missing copyright attributions on some really old BSD code. I see trouble ahead.
Maybe they're like Sisyphus, and for some great cosmic wrong, SCO has been condemned to try one legal theory after another. Just as they near the top of the hill and are about to make a buck at last, down they fall and then have to start up the hill again, lugging a new and equally doomed legal theory. It's kind of heart-wrenching to watch.
Let's take a look at the viability of their latest claim.
I at first thought I shouldn't write about UNIX history until I understood it. Then I realized that nobody on planet Earth understands UNIX history. It's unknowable. At least by normal mortals.
There are too many pieces and too many missing pieces, but I'll have to try to muddle through, like poor SCO will at trial, if they actually do sue someone for failure to "reinsert" the copyright notices. I'll just plunge in, therefore, and you can watch my UNIX education unfold before your very eyes. How's that for daring and panache? All you UNIX greybeards can chime in any time and correct any errors you see.
It helps that Groklaw has come into some of the documents from the BSDi trial. But first, before we get to that, what is it SCO is now claiming?
SCO's BSD "copyright infringement" claims are a bit clearer, thanks to an article in PCPro. It shows that SCO understands the BSD case about as well as it understands the GPL. Here is Stowell's characterization:
"SCO's PR Director Blake Stowell told us that a case between AT&T and Berkeley Software Development (BSD) was settled where the latter had to remove certain files and reinsert copyrights that had been stripped out. 'Some of these same files have now gone into [the Linux kernel] (both those that had to be removed because they were never supposed to be in BSD in the first place, and the files for which copyright attribution had been stripped away. The copyright attribution has still been stripped away as they were contributed into Linux). This is a violation of SCO's copyrights,' he said.
'We haven't yet gone down the path to determine how the code got into [the Linux kernel]. We only know that the code is in there... the problem really impacts the Linux business user more than anyone else, because all liability rests with them,' he added."
So, selective enforcement only against deep pockets. Let's go down the path to determine how the code could have gotten into the kernel, if in fact it has, something not yet proven. First, it seems AT&T didn't have any copyright notices on their files until 1992, because they didn't want any copyright notices on their files until then. They first applied for copyright for 32V in 1992. By that time, they had been licensing the code for more than a decade. 32V was first published in 1978, as you will see in the judge's opinion. That's one reason they had little hope of prevailing in the BSDi lawsuit. Here is what the judge said about their copyrights, and, yes, it's long, but UNIX history is like that:
"In order to prevail, Plaintiff must prove that it has a
valid copyright in the UNIX code. Plaintiff's chief difficulty
here is the 'Publication doctrine.' The publication doctrine
denies copyright protection to works which the copyright owner
'publishes,' unless the owner has properly affixed a notice of
copyright to the published work. This doctrine has suffered steady
erosion over the years, and it now applies in full force only for
works published prior to January 1, 1978. For works such as 32V
(published in 1978), which were published after that date but
before March 1, 1989, the doctrine is subject to the escape
provisions of 17 U.S.C. Section 405(a) and the common-law 'limited
publication rule.' For works published after March 1, 1989, the
publication doctrine has been eliminated by the Berne Convention
Implementation Act, 102 Stat. 2857 (1988).
"Because Plaintiff copyrighted 32V in 1992, Plaintiff
benefits from a statutory presumption of validity and BSDI has the
burden of proving otherwise. Williams Electronics, Inc. v. Arctic
International, Inc., 685 F.2d 870, 873 (3rd Cir. 1982). BSDI seeks
to carry this burden by proving that AT&T and Plaintiff have widely
published 32V without proper notice.
Publication is defined as:
the distribution of copies or phono records of a work
to the public by sale or other transfer of
ownership, or by rental lease or lending. The
offering to distribute copies or phonorecords to a
group of persons for purposes of further
distribution, public performance, or public display,
constitutes publication. A public performance or
display of a work does not of itself constitute
publication.17 U.S.C. Section101.
"Version 32V source code has now been distributed,
without notice, to literally thousands of licensees.
Consequently, Plaintiff can have no valid copyright on 32V unless
it can fit within one of the statutory or common law escape
"The three statutory escape provisions are listed in
section 405(a). These provisions relieve a copyright owner from
the harsh consequences of noticeless publication if the owner
(i) omitted the notice from a 'relatively small number of copies;'
(ii) registers the work within five years of publication, and then
makes a 'reasonable effort' to add notices to the noticeless copies
already distributed; or (iii) proves that a third party omitted,
notice in violation of an express agreement in writing 17 U.S.C Section
"Plaintiff cannot avail itself of any of these provisions.
Notice was omitted from thousands of copies of 32V; no contractual
agreements require the licensees to affix notice;
Plaintiff failed to copyright 32V until 1992, well over five years
after 32V was published; and Plaintiff has not yet made reasonable
efforts to add notices to the many noticeless publications of 32V.
Consequently, Plaintiff must try to fit within the common-law
doctrine of limited publication.
"Under the common-law doctrine of limited publication, a
publication will not divest an author of copyright protection if
the author "`communicates the contents of a manuscript to a
definitely selected group and for a limited purpose, and without
the right of diffusion, reproduction, distribution or sale. . ."
Rexnord, Inc. v. Modern Handling Systems, Inc., 379 F. Supp. 1190,
1197 (D. Del. 1974) (quoting White v. Kimmell, 193 F.2d 744, 746-47
(9th Cir. 1952), cert. denied, 343 U.S. 957 (1952); see also
American Visuals Corq. v. Holland, 239 F.2d 740, 744 (2d Cir. 1956)
(quoting White v. Kimmell). There is some question as to whether
the doctrine of limited publication survived the 1976 amendments to
federal copyright law, but the weight of authority holds that its
has. See Nimmer on Copyrights, Sect. 4.13[B](1992), and the cases
cited therein ('the doctrine of limited publication continues under
the current Act').
"For the present dispute, the key criterion is that the
alleged copyright owner must only have distributed the works to a
'Selected group.' To qualify as a selected group, those receiving
the work must be more than just customers self-selected by their
desire to purchase the work. Otherwise, 'all the purchasers of the
work' would qualify as a 'selected group,' and all publications
would be limited publications. For example, if a programmer sells
software to anyone willing to meet the asking price, then the
programmer can hardly be said to have communicated the work to a
selected group. See Brown v. Tabb, 714 F.2d 1088, 1092 (11th Cir.
1983) (the group containing all of the potential purchasers of a
jingle was not a 'selected group'); Academy of Motion Picture Arts
v. Creative House Promotions, 944 F.2d 1446 (9th Cir. 1991) (Oscar
winners are not a selected group because Plaintiff did not
distribute its Oscars generally to the public). Situations where
courts have found selected groups include distribution of
manuscripts by authors to a small collection of friends; of floor
plans to contractors so they can prepare bids; of advance copies of
works to colleagues for criticism or review; of samples to
potential customers; and of promotional material to dealers.
Himmer an copyright, Sect. 4.13[A] (1992).
"As the above examples suggest, the traditional scope of
the limited publication doctrine is narrow. But courts have shown
a tendency to expand the scope of this doctrine when applying it to
computer products. See Nimmer on copyright, Sect. 4.13[A], [2).
For example, several courts have declined to find a publication
where a copyright owner has communicated support materials to
customers buying computers. One such case is Hubco Data Products
Corp. v. Management Assistance Inc., 219 U.S.P.Q. 450 (D. Idaho
1983), where the court faced the issue of whether the plaintiff
computer manufacturer had published the operating system
distributed with its computers. The computers (complete with
operating systems) apparently were available to anyone willing to
pay their price, but the court none the less found only a limited
publication of the operating system. The court reasoned that the
'operating systems were only offered to owners of [plaintiffs)
computers.' Id. at 455. See also Data General Corp. v. Digital
Computer Controls, Inc., 188 U.S.P.Q. 276 (Del. Ch. 1975)
(schematic drawings were the subject of a limited publication
because they were only distributed to those who bought computers).
"The limited-publication variations endorsed in Hubco and
Data General, even it valid, are of no help to Plaintiff here.
Version 32V was not distributed as an ancillary communication
supporting the sale of a computer -- 32V was the actual product
itself. Consequently, the only question is whether AT&T and
Plaintiff limited 32V's distribution or offered it to whomever
"Plaintiff argues that it has only distributed 32V to
three select groups of licensees: educational organizations, U.S.
government agencies, and carefully-screened corporate entities.
Although the former two groups could conceivably qualify as
'selected,' the latter can qualify only if the screening process is
suitably restrictive. According to depositions of Plaintiff's
personnel, the screening includes investigating the kind of code a
potential customer seeks (source or binary); detailed information
about the nature of the customer's business; the customer's
financial ability to pay; and whether the customer is in a country
that protects trade secrets. Plaintiff needs all of this
information to assess three features of its prospective customers:
which product is best for the customer, whether the customer can
pay, and whether the customer will take good care of Plaintiff's
intellectual property, (Kennedy Decl., Ex. C at 1622, Ex. D at
111-12; Pl.'s Reply Br. at 10.)
"Even accepting this description of Plaintiff's screening
process as true, it is hard to see how the screening would yield a
selected group of corporations within the meaning of the doctrine
of limited publication. Plaintiff essentially seeks what any
lessor of property seeks: someone who will take good care of the
property and pay the bills as they fall due. If this were enough
to create a selected group, then practically anyone who leased
property would belong to a selected group.
"The copyright laws in effect prior to 1989 do not allow
such expansive protection against the consequences of noticeless
publication. As quoted above, publication includes distributing
copies by 'renting, leasing, or lending.' 17 U.S.C. Section 101. Under
Plaintiff's construction of the doctrine of limited publication,
'renting, leasing, or lending' would not constitute publication
whenever the owner of the copyright screened out irresponsible
customers. It seems unlikely that Congress drafted this section
with this purpose in mind.
"Consequently, I find that Plaintiff has failed to
demonstrate a likelihood that it can successfully defend its
copyright in 32V. Plaintiff's claims of copyright violations are
not a basis for injunctive relief." [emphasis added]
So, if they had no claim for copyright violations, we need an explanation as to why copyright notices were to be put on these files, when the judge said they had no likelihood of being able to prove they had any such rights. According to Blake Stowell's remark, the answer is that sometimes you can get in a voluntary settlement what you can't get from a judge. And evidently that is what happened. At the time of the voluntary settlement, the parties agreed to put copyright notices on certain files. This is more than the judge seemed willing to grant. Stowell's characterization of copyright notices having been stripped off and needing to be reinserted, which makes it sound like Linux or somebody did something wrong, seems misleading.
PC Pro interviewed an attorney who clearly feels SCO's position is very weak, because any "infringement" is de minimis anyhow, even in a best-case scenario for SCO, and then he says this about the settlement:
"Open Sources: Voices from the Open Source Revolution , O'Reilly (ISBN 1-56592-582-3) contains a potted history of the case in question, and describes the settlement agreed on 4 February 1994 thus: 'The result was that three files were removed from the 18,000 that made up Networking Release 2, and a number of minor changes were made to other files. In addition, the University agreed to add USL copyrights to about 70 files, although those files continued to be freely redistributed.'
"'Moreover,' says Schultz. 'if SCO then released its own Linux distribution under the GPL, that would presumably trump any previous restrictions they made on BSD, since the Linux distribution was more
recent. It would certainly seem reasonable to me to assume Linux was legal if SCO released their own version under the GPL, which it appears they did.'"
So they hope to make money from 3 files that were supposed to be removed and they claim were not and some copyright notices on maybe 70 files that were perfectly legal to use and that SCO, in its Caldera persona, later released under the GPL? Somebody send poor Mr. Skiba a memo, will you? I think he may need to adjust his math. And where, in this picture, are the damages that SCO has suffered? Here's what the judge in the BSDi case wrote about the amount of code overlap at issue in that case:
"The alleged overlaps
between parent and child probably amount to less than a percent of
the total. . . . Indeed, ignoring header
files and comments (see below), the overlap in the critical
'kernel' region is but 56 lines out of 230,9995, and the overlap
elsewhere is 130 lines out of 1.3 million."
Compare what Stowell said, that the copyright notices were stripped and were supposed to be "reinserted", with the answers given in USL's responses to BSDi's second set of Interrogatories from USL v BSDi, Interrogatory 6, one of the documents we now have obtained:
"Identify any file in the UNIX/32V source code that at anytime has been distributed, whether pursuant to license or otherwise, without a copyright notice.
"All files contained within the UNIX/32V have been distributed by AT&T or USL without a copyright notice pursuant to the license agreements referenced in response to interrogatory number 1, all of which restrict the licensees right to make copies of such source code and limit the distribution of derivative works to other UNIX/32V source code licensees."
What I gather is that the code never had a copyright notice in the first place because they chose trade secret over copyright protection. You couldn't have both, so they removed any copyright notices themselves and distributed the code for years that way. Here is their answer to Interrogatory No. 5:
"AT&T and USL have distributed the UNIX/32V source code without a copyright notice pursuant to the license agreements referenced in response to Interrogatory No. 1, all of which restrict the licensee's right to make copies of such source code and limit the distribution of derivative works thereof to other UNIX/32V source code licensees."
So, how am I doing so far? I'm getting that there were no copyrights on this code originally because AT&T deliberately removed any such notices, so they could rely on trade secret status instead. Nobody stripped off their copyright notices against their will or did it to steal IP and obfuscate the theft. They did it themselves. So for years, the files were distributed without copyright notices because that was their free will choice. Then in 1992, they filed for copyright, but the University challenged it, and the judge's opinion was that they had no valid copyright. If SCO really wants to retry the BSDi case, I hope someone notices that detail. I'm sure they do.
Apparently, then, the super hush hush agreement that settled the case required a few files to have copyright notices put on them, according to what the attorney quoted above is saying. No one knows for sure until we see the sealed agreement which files, from all I have discovered so far. That was to be done by the parties to the agreement. Linux was not a party to that agreement. If there was a failure to follow through, who would you sue? Linux?
Keep in mind that Linux began in 1991. So before the trial and all while the trial was going on, code was being donated to Linux that everyone believed was legal, and it was legal, from what I am seeing. Then, when the trial ended with a settlement, the parties made a secret agreement between themeselves that from that day forward, they'd put copyright notices on a handful of files. But those files may well have been in Linux already. And Linux wasn't party to the settlement and would have no way of knowing what files were to have copyright notices attached unless the parties revealed it. Wouldn't that seem to be a logical explanation for any lack of copyright notices in those source code files, if in fact such proves to be the case?
And what would be the remedy here? How about telling everyone what files are involved so the copyright notices can be put on the files? Is that too simple? Not monetized enough? There was reportedly another clause in the agreement, namely that USL was never to sue anyone using 4.4 BSD-Lite as the base for their system.
See what I mean about a curse? How are you going to make money from this mess? It gets worse for SCO. It seems, reading the judge's opinion and the interrogatory answers, that a lot of trade secrets got spilled here and there along the way too. That's not good for SCO. A trade secret is only a secret for as long as it's a secret.
There were thousands of UNIX licensees worldwide that either had, or were eligible to receive, the UNIX/32V sources or it's derivatives. Some of them were individuals, not entities. It can only be an estimate, because the bookkeeping wasn't stellar. Here is how AT&T and USL described their own licensing records at the central licensing records repository, from the same document, USL's responses to BSDi's second set of Interrogatories:
"USL maintains original license agreements and related correspondence at a central repository located in Greensboro, N.C., which generally contains correspondence related to licenses, amendments of modifications to such licenses and correspondence with licensees. . . .Included among such documentation are electronic data bases which comprise USL's best, although incomplete, list of such information, and which are known to contain inaccuracies." [emphasis added]
"Known to contain inaccuracies." And "incomplete." That's how well the trade secrets were guarded. There were, then, thousands of copies of 32V distributed, but the best records about their disposition are inaccurate and incomplete. According to several depositions in the case this best record didn't reflect the fact that the University of California at Berkeley itself even had a UNIX/32V license. Talk about your incomplete. Some of these UNIX trade secrets may not have been too closely tracked or guarded due to the sheer number of unmarked files, records, or documents that were simply lost, misplaced, or otherwise unaccounted for. The UC Regents Amicus brief, footnote 24:
"USL produced a 1986 data base of its source code licensees,
bates nos. P8832-9312, which is too large to attach in total.
However, the data base contains approximately 5,760 licensees for
AT&T's various software products. A large number of these licensees
could obtain 32V though AT&T's source code exchange program.
Frasure Depo., at 54:25-57:8. The voluminous data base does not even
list every source licensee for 32V, as evidenced by the absence of
any 32V entries for the University of California."
And they would like to claim a trade secret and copyright violations on this pile of spaghetti?
Let me see if I can summarize. The sealed settlement agreement was a big secret, with no one allowed to tell what the terms of the settlement were. The trade secrets were spilled all over the highway, so to speak. They would now like to hold Linux enterprise users responsible for not abiding by the terms of this secret agreement. They expect to collect money from people not party to a secret agreement for failure to live up to the terms of the secret agreement to which they were not signatories? Um, a lot of money? How? There is one more pesky problem from SCO's point of view, as the attorney pointed out. Caldera open-sourced UNIX/32V and destroyed any trade secret protections it had previously enjoyed.
This isn't a failure in the open source development model. This crazy code is like a bucking bronco. They lost control of it every which way. It's a failure of AT&T's and USL's distribution model.
For a company wanting to get rich quick, I foresee a problem with their new legal theory, but it'll be worth standing by and watching SCO trudge up the hill with it. Like I say, maybe it's a curse. I think they should maybe just give up and have no further truck with UNIX. Just slap a GPL on it and call it a day. Maybe then the curse will lift and they can find some peace.
Here are some handy links that will help you get up to speed on the BSDi case. No reason I should suffer all by myself: