Now that SCO has sold off the caldera.com domain name, their previous robots.txt file no longer blocks access to the legacy Caldera web pages on Internet Archive. And what has popped up?
Some evidence that I believe proves that the SCOsource licensing program in 2004 was a right to use SVRX code, which would make it code SCO has to pay royalties received to Novell and which it must ask Novell's permission to license under the terms of the APA. That is not what SCOfolk testified to at the bench trial in 2008, where they presented the SCOsource license as protection from litigation or in one case as shared libraries. So this is a new piece of evidence to add to the pile that indeed at least some of the SCOsource licenses were right to use licenses for SVRX code.
Here's the page that now resolves properly on Internet Archive: SCO's SCOsource License Program page. The page describes the licensing offering that was SCOsource as of 2004:
Since the license pertains to SCO IP that the end user already received in the unauthorized Linux distribution, the SCO license doesn’t include a media kit. See? SCOsource was not just about buying SCO off so they wouldn't sue you. The conceit was that Linux users had already received the code, albeit improperly in SCO's twisted thinking, so no code needed to be distributed with the license. They already had it.
Why is that important? Well, for one thing, Judge Dale Kimball ruled otherwise:
Separate from its licensing of products, SCO began entering into SCOsource licensing agreements that were unique in that they did not involve product. Instead, these license agreements were waivers and releases of conduct based on the buyer's use of Linux. I believe now, based on the resurfaced Caldera web page, that he erred when he ruled that those SCOsource licenses were chiefly about avoiding litigation and not about product. He also ruled:
The central features of the SCOsource agreements are a covenant not to sue and a waiver of claims by SCO for the buyer's internal Linux usage. The agreements grant rights and a license to use SCO IP. "SCO IP" is defined as SCO UNIX-based code, and "UNIX-based Code" is
defined to mean "UNIX System V or UnixWare, or (v) any modification or derivative work based on or licensed under UNIX System V or UnixWare." The purpose of these licenses was to excuse the licensee's purported infringement of SCO's IP. I think that was error too, first because the issue should have been what SCO intended to sell, and at the time it absolutely did claim to own Unix and UnixWare; and second, because I believe the resurfaced web page proves that they meant Unix alone, as a product being licensed, in at least one SCOsource offering in 2004. I'll show you the detail that I believe proves it.
SCO IP is defined in the agreements to include UNIX System V, which would appear to include SVRX. But given the terms of the amended APA between Novell and SCO, as this court has previously ruled, the SVRX copyrights did not transfer to SCO. Therefore, SCO IP cannot include SVRX and can only mean SCO UnixWare....
Because the SCOsource licenses cannot be construed to include a release of SVRX copyright infringement, the court does not find the licenses to be SVRX Licenses that generated SVRX Royalties to Novell under the APA. SCO could not release Novell's rights to claims based on its ownership of the SVRX copyrights. Even if the releases contained in the SCOsource Agreements were considered SVRX Licenses, there is no value in the agreements with respect to Novell's SVRX interests. As such, Novell has no entitlement to monies SCO received with respect to a release of only SCO's rights. The value of those SCOsource releases is a matter between SCO and the parties who entered into such releases. In addition, because the court concludes that the releases in the SCOsource Licenses were not SVRX Licenses, SCO had authority to execute the agreements.
Did you notice the registered trademark symbol after the word UNIX? It's a dead give away, to me, that this is *not* UnixWare being discussed by SCO in 2004, because UnixWare has its own registered trademark, Serial Number 78833448. Unix is Serial Number 73544900, which you can confirm by doing your own search on USPTO.gov. That doesn't match SCO testimony either, of course, since they went on and on about how UnixWare and Unix were the same thing, tree trunks and branches, blah blah. Evidently, the USPTO doesn't agree that they are the same at all, and neither does X/Open Group, which owns the trademarks. And neither did SCO in 2004, when SCOsource was young, judging by its use of the UNIX trademark symbol. That trunk-is-the-tree story came later.
'Right to use' is a term you see in normal licensing of software. For example, notice the use of the term in IBM's 1985 agreement with AT&T in the Grant of Rights section:
2.01 AT&T grants to LICENSEE a personal, nontransferable and nonexclusive right to use in the United States each SOFTWARE PRODUCT identified in the one or more Supplements hereto ... And here's the wording in EV1's SCOsource license, in its Grant of Rights section:
2.1 SCO hereby grants Licensee and Licensee hereby accepts from SCO the following limited, non-exclusive, non-transferable right and license to use SCO IP ... They are both, by my reading, offering a right to use software, meaning product. SCO's definition of "SCO IP" in that license reads, "1.10 'SCO IP' means the SCO intellectual property included in its UNIX-based Code in Object Code format licensed by SCO under SCO's standard commercial license." There were two versions of the EV1 license, which you can see compared here, but they both offer a right to use code. So despite SCO's squirrely language, it's talking about code, a right to use the code being licensed, the only difference being that normally one gets the license and code at the same time. SCO's offering was a licensing after the code allegedly was already in hand. If the various restrictions on use of the code were obeyed, *then* in section 3.0, SCO promised not to sue you. But the core of the license was the right to use the code.
Kimball's conclusion led him to rule that SCO didn't owe Novell any of the money it took in from the EV1-style victims or need Novell's permission to sell a license to folks like EV1, in that he had ruled they couldn't be SVRX licenses, in the Final Judgment:
4. With respect to Novell's claims for Declaratory Relief (Count IV), pursuant to the Court's Memorandum Decision and Order dated August 10, 2007, Novell is entitled to direct SCO to waive claims against IBM, Sequent, and other SVRX licensees; Novell is entitled to waive such claims on SCO's behalf; and SCO is obligated to recognize such a waiver. In addition, pursuant to the Court's Findings of Fact, Conclusions of Law, and Order dated July 16, 2008, SCO was not authorized under the APA to amend, in the 2003 Sun Agreement, Sun's 1994 SVRX buyout agreement with Novell, and SCO needed to obtain Novell's approval before entering into the amendment; but SCO was fully authorized under the APA to enter into the 2003 Microsoft Agreement and the SCOsource Agreements with Linux end-users without any approval by Novell. But as we've just seen, SCOsource, at least in 2004, *had to be* about Unix SVRX. I think it can be easily discerned, actually, by what he also wrote about Darl McBride's testimony in the Final Judgment:
Darl McBride, SCO's CEO, testified that during his conversations with Greg Jones at Novell he pointed out that SCO's efforts to enforce the intellectual property in Linux would indirectly help the sale of the various UNIX flavors that compete with Linux in the market place and that such a boost would potentially increase the declining SVRX Royalty stream that SCO remitted to Novell from contracts that licensed out the older products. That would only be true if SCOsource was about SVRX, after all, since SCO owned Novell nothing for UnixWare licenses.
Why does it all matter? It's significant in that if SCO were to persuade Judge Ted Stewart to hand over the copyrights after all, they'd have a green light to sue end users some more.
Kimball believed SCO's testimony about SCOsource, I guess. If you recall, on day one of the bench trial in 2008, Chris Sontag, who was in charge of running SCOsource, testified that the license was several offerings, but that it was basically to protect you from litigation, that it wasn't about SVRX code, only UnixWare and shared libraries.
SCO had blocked access to the evidence Novell could have used to demonstrate the truth of what SCOsource really was by removing pages and blocking access to them on Internet Archive. Now the evidence is out in the sunlight.
And just in case it disappears again, here are more relevant sections of the page:
Here's the explanation for the pricing, and then the part about why you didn't get code with the license, as you would in a normal license deal:
Notice the pricing structure, depending on what you used Linux for and how many CPUs were involved, which would make no sense if what you were buying was simply protection:
As you can see, it absolutely was about code, but about code users had already received that SCO was now making "legal" to use. This doesn't match Sontag's testimony, to the extent I understand it.
So. There. You. Are.
In the old days, it seemed every time Groklaw linked to evidence on one of those Caldera pages, it went to the great 404 in the sky within a day or two. Now, it will be possible to fix all those links. I wish they'd sell the sco.com domain name next. Maybe then we could get the complete historical picture.