Meanwhile the SCOfolk are already spinning to the media. McBride is all about appealing the August 10 ruling, according to the Salt Lake Tribune's, to me, funny account:
I don't want to burst any bubbles, but that is exactly what he just had. The appeals court will be looking at the law, not the facts. SCO lost, fair and square.
First, the courtroom is lovely: nice woodwork floor-to-ceiling on the walls; ornate plaster ceiling in off-white with blue trim and (if I remember right) gold accents. Not a bad place for Judge Kimball to spend his time.
On to the trial... my comments are paraphrases of the testimony. I used quotation marks when my text is verbatim quotation of what was said (or near verbatim subject to my memory and/or legibility of my notes).
SCO called their last witness, Mr. Nagle, Senior Director of the Product Group (OS and mobility). He has an engineering background going back to '84.
Mr. Normand, I think, was doing the direct.
Q: relationship between UnixWare vs previous Sys V?
A: UnixWare consists of SVR4, is in Unixware "lock, stock, and barrel" plus some NetWare enhancements and user interface additions. Largely the same as SVR4.2
Q: (showed diagram with two circles, mostly overlapping; both of them mostly contained within a larger rectangle, part of the circles descending outside the rectangle at the bottom)
In the diagram, UnixWare 2 is shown as the rectangle. The circles represent SVR4.0 and SVR4.2MP. The part of the circles outside the rectangle is labeled as having no commercial value. I don't recall the exact question posed in conjunction with the showing of the diagram.
A: 4.2MP is in UW2 "lock, stock, and barrel"
Q: Would UW operate without SVR4?
Q: Asking about UW 2.1 and later
A: 100 engineers spent several years to develop 2.1. There was industry cooperation, too, what we call "data center acceleration."
UW 7 combined Open Server features.
In general we "improved it, increase value in the marketplace."
Q: Do you know what is meant by the term syscalls?
A: Gave the usual description, background on user-level vs kernel-level, "trap" instruction to request kernel services from user mode, syscall being the mechanism for passing information between user and kernel.
Q: Compare syscalls between SVR4 and UW
A: "one and the same"
A: Usual description: input/output, communication with hardware devices
Q: SVR4 Streams in UW?
A: Description about elf executable format and dynamic linking support going together.
Q: ELF in SVR4 and UW?
Q: Memory allocation?
A: Typical description
Q: Compare memory allocation in SVR4 and UW
Q: File System?
A: Described variety of filesystems
Q: Compare SVR4 and UW filesystems
Q: Familiar with Sun Solaris? How was it developed?
A: At time of SVR4, Suan announced (intent to) support, made agreement with AT&T; some features came from Sun; latest Solaris is based on SVR4.
Q: How do you know these things?
A: Familiar with the affinity between the two systems, e.g., both use ELF. Moving applications between the two OSes is straightforward.
Q: Was knowledge of Solaris part of your employment?
A: I tracked the cooperation between the systems. ... It was widely known that Solaris used SVR4.
Q: Another diagram, similar to the first but with an arrow leading from the SVR4 circle inside the UW rectangle and over to a new circle labeled Solaris. (Showing that Solaris is based on SVR4.)
A: Solaris is based on SVR4 and on BSD, Sun OS (previous Sun UNIX version), and other sources.
Q: Is SVR4 in Solaris?
A: Yes to the best of my knowledge
Q: Does it have Streams?
A: Yes to the best of my knowledge
At this point Mr. Normand ended direct.
(I didn't catch the name of the Novell attorney doing cross.)
Q: (Referring back to the first diagram shown during direct) went through his understanding of what the diagram represents. (If I remember right, it was at this time that it was pointed out that the SCO-developed IP is the part of the rectangle that's outside the circles.)
Q: (Showing the second diagram, which includes a circle for Solaris)
Q: There is no arrow from UW to Solaris (only from the SVR4 circle to Solaris). There is no code unique to UW in Solaris?
A: Not that I'm aware of
Q: Do you agree the graphic suggests this?
Q: How many lines of code (LOC) in the top/bottom parts of the circles? (In the diagram, the top parts in included in UW while the bottom parts are not (and are marked "no commercial value").)
A: I don't know
Q: Haven't you gone through it?
A: I have not.
Q: (along the lines of) So you are just assuming it has no value?
A: Not just an assumption. (Though I didn't catch anything in his response that explained the basis)
Q: You can't answer line by line.
Q: How many total LOC (in SVR4)?
A: I don't know
Q: Greater than 1 million?
A: Almost certainly
Q: Greater than 2 million?
Q: 5 million?
A: I don't know.
Q: So it's probably between 2 and 5 million? Agree?
A: No. It might be more than 5 million.
Q: What proportion / Does the diagram show proportionality of the part of the circle inside UW rectangle vs below it?
A: Diagram doesn't reflect proportion.
Q: Was it intending to convey that a "small percentage was left out"?
A: It was intended to convey that a "small percentage" was left out.
Q: (I didn't catch it exactly, something about) you don't know; it could be greater
A: Correct, it could be.
Q: Estimate? 10%?
A: I would be surprised if it were greater than 5%
Q: What about code that was changed from SVR4 to UW? Would that be counted as above or below the line (of the bottom of the UW rectangle)?
A: If changed, it would be considered part of the OS
Q: (something about ELF)
A: I don't know. Speculate might be some additions. (Sorry, I don't understand my own notes here.)
Q: SVR4.2 in UW2? (proportion?) You don't know here either?
Q: Would it be fair to say the lower portion might be hundreds of thousands of LOC?
A: It's possible.
Q: Open Solaris. Do you know about it?
Q: Given Solaris 4.0, would Open Solaris be expected to contain significant SVR4 code?
A: Yes, could be
Q: Is there commercial value to this code?
A: Not if it doesn't contribute to the customer choosing it.
Q: Assume 1 million LOC: 900K owned by you, 100K owned by someone else. The owner of the 100K says you need to remove it. I would think it would then have substantial value!
A: Not if it's not essential.
Q: You have 20 years of experience in this area. Your testimony is that it would be trivial to remove 100K LOC?
A: Not what I'm saying. We might be able to remove those lines of code.
Q: We can agrre it's not trivial to remove 100s K LOC?
A: Depends on the subsystem. "Some is easier to remove than other. Others would be difficult."
Q: Are you familiar with the OpenSolaris license, that the public can see the code?
Q: It's the 2003 SCOsource license that gave Sun this right to expose SVR4 to the public?
A: The right to expose UW code also.
Q: It gave the right to expose SVR4 code?
Q: The right to open code has value?
A: That's a debatable point. There are "those who would say ... [it] ... does have market value" such as Sun. Others argue there's no market value, that protected code has more value than open code. I concede it might have market value.
Q: Referring to deposition p 26 lines 24-27 -
Right to release code is something that has market value?
Q: Is that your testimony?
Q: You agree it has value?
Referring to deposition, bottom of p24, top of p25
Q: agree it has market value?
A: lots of debate: "you don't pay to be able to read it"; marketing value to a company to be perceived as "open"
Q: Is every line of SVR4.0 in UW?
A: Most likely not
Q: What about 4.2MP?
A: Highly probable yes
Q: Vast majority of 4.0/4.2 in UW?
A: Yes absolutely
Q: Was the graphic (referring to the diagram discussed above) meant to depict that?
Novell said they had some evidentiary matters/submissions to clear up, should we do that before or after closing?
There was some confusion as to who would close first. Novell expected to but docket said SCO would be first. Attorney for SCO (Singer?) said he was ready if he needed to go first. Novell proceeded first.
More to come, after he reviews more notes.
Okay, picking back up where I left off...
Mr. A speaking for Novell (I didn't catch the actual name. It looks like there is an Acker, that could be it.)
The closing seemed to go on for quite a while, a far cry from what you see in movies/TV and what I've seen in traffic court. :) Fortunately, they give a binder to the Judge with all they information they presented. Judge Kimball listened attentively, of course. My notes on the closings are not as detailed as the testimony (I was paying more attention to Judge Kimball's face as the closing arguments were delivered. He really didn't give away anything through his reactions, though.), but here's what I have on what was covered in his closing statement. Novell's closing:
Let's get back to the basics of the case.
APA Schedule 6, 1.1A, Exhibit 1: SCOT has to remit royalties to Novell.
1. How much of MS/Sun money goes to Novell? The court already determined that SCO breached its fiduciary duty. The question is how much must they pay to Novell.
2. Sorry, I missed this one.
3. Was SVRX licensing "merely incidental"?
Going back to 2002, Mr. McBride and Mr. Sontag told us SCO's financial situation was horrible.
(Cites deposition: losing millions; in bad shape.) McBride was asked what he would do. He said he believed there was UNIX IP in Linux. Most of SCO's revenue was from UNIX. The key strategy identified: Market the UNIX assets. Revenue from UNIX sales was "marching South."
They intended to "maximize the value of the trunk" (referring to the diagram of the tree with SVRX as trunk, UW as branches.)
What is it Mr. McBride decides to do?
Sontag (from deposition): Asked if SCOsource was to generate billions, A: potentially significant. Q: Billions? A: "potentially yes"
Again Sontag talked about the "trunk": Q: "entire body of IP?" "That was the plan."
This is the tree. Instead of "branches," "Mine the trunk!"
Mr. McBride's words on "the trunk" in a deposition p265 lines 6-13: Q: Sun/MS: referred to [trunk], correct: A: "that's the way I'd" look at it.
Next thing we need is lawyers. Not a sale; licensing! Need litigators! "You have to pay us so you can run Linux" and be OK with SCO. From depositions: SCO license to "make Linux users 'clean' with SCO" (in case of inadvertent use of SCO IP). From deposition of Mr. Petersen, cited documentation of litigation.
If it's about selling, why the need for litigators? It's about a licensing scheme.
Mr. Hunsaker made this clear in his email (2003?): "There is no connection between Linux and OpenServer"
Dec. 2004: Let end users know if they want to use Linux safely they must pay. "1500 payments" expected.
SCO stated it "holds the rights to UNIX" (UNIX emphasized) not "to UnixWare." "Get in line, get your license, or you're gonna get sued."
It's in this context that Sun and MS deals were executed. They were SCOsource licenses.
It can't be disputed that the Sun license is a "restatement" of the SVRX license. Sontag was upfront about this. (In deposition, which is cited.)
There's no question they (SCO) allowed Sun to open-source Solaris.
Is there market value in that (the right to open source Solaris)? SCO would have you believe "no." Their own engineers testify "yes." Solaris was 'frozen in time'" -- older code: the trunk of the tree. Cites the "it would not be favorably received" line from earlier in the trial and the humor in that notion.
Also, SCO was concerned that (citing Pattersen deposition) Sun could use its expanded rights to protect its Linux users.
That would have commercial value! Extremely valuable, it could seriously undermine SCOsource. Under the 2003 license, Sun could now indemnify its Linux customers: substantial commercial value!
Testimony in response to the question "is all prior code in UnixWare", the answer is they don't know!
There was no evidence that open sourcing Solaris via SCOsource license to UnixWare covers all Solaris code. Nobody did the analysis.
Solaris was based on SVRX.
"For SCO to argue there's no commercial value simply defies logic and ignores the testimony of their own current and former employees."
(I missed some stuff here. I'm not certain the following reasons relate directly to the above or to a point brought up in between -- perhaps the idea of the licensing being merely "incidental.")
Reason 1. There is no question that this is not a routine license. It's SCOsource.
2. Practice here is inconsistent with standard SCO licensing procedure. (Refers to license listing only current versions and not past versions.) (Comment on suggestion that it was just a "sales technique" to only use the term UnixWare (rather than listing SVRX.)
They list 30 other prior pieces of software!
3. Standard SVRX license terms: no derivatives, no open source (their own witness told us that)
4. UnixWare revenue is booked as Unixware. (Deposition or testimony of SCO accountant) But this wasn't! Per SEC filings, this was SCOsource. "Mr. McBride said he would never make a mistake in an SEC filing." (BTW, that particular part of the testimony was one of my favorite parts of the reports from earlier in the trial. Too bad I wasn't there that day to see it live. As I read it, I saw the trap being set and Mr. McBride taking the bait.)
5. Sun already had a license for the legacy software. The only reason to list prior pieces of software was to expand the confidentiality provision.
6. SCO says they've lost hundreds of millions of dollars [due to Novell contesting ownership of the SVRX IP]; now they say it had no value.
"Oh no, no, no, no, no, no! No value. ... Incidental."
7. SCO's position now is inconsistent with 2007 counsel response when Novell requested to see the license. And to audit. Did they (SCO) argue these weren't SVRX licenses? No.
(My notes are little unclear here. Maybe someone who better knows the context and citations can make sense of it.)(Cites Mr. McBride testimony/deposition "I would look to Tibbitts letter." Tibbitts letter: "we haven't finished the audit. By citing the APA, it appears you are concerned about royalties." License at time of APA does not extend to 2003 Sun or to new(!) MS license. From deposition/testimony: Q: In 2003 was the reason that it was after APA? McBride: Yes.
"That was then. New defense now."
The Microsoft deal:
Sections 2 and 4, royalties to Novell.
Section 2: states all of SCO's IP
Now: they were only concerned about the new stuff.
Previously: Mr. McBride: yes they were concerned about SCO IP including SVRX. "That was the license..." UnixWare and older UNIX technology. 1.5 million. None was proided to Novell.
Section 4: 8 million. Exhibit C: UnixWare plus a long list of prior IP. New rights included sublicensing.
"It's the trunk of the tree."
"Incidental" should be rejected for the same reasons as with the Sun agreement.
Then: "100s of millions of dollars"
Other SCOsource licenses, including Everyone's Internet: "license for SCO IP": "UNIX System 4 or UnixWare." (I suspect "System 4" actually means "System V Release 4.")
"Mining the trunk of the tree."
This, too, is revenue that should have been passed to Novell.
Concede 7 million of MS license is just UnixWare.
Commented that this understanding should have been reached in 2003. Even now they don't provide any apportionment.
It is telling that Mr. McBride says this about the Sun and MS deals: "my view of these 2 licenses is that Novell had no more standing than the court recorder" to ask for copies of the licenses.
Novell asks for 9 million (and more digits to the right) for Sun deal, 9 million (and more digits) for MS deal, and for the other SCOsource deals, we ask for all of it. 19 million total (and more digits of course).
Justice would not be served if SCO were to [get away with it (I missed how he put it, sorry)].
Mr. Singer, I think it was, argued for SCO.
This court ruled a SVRX license was implicated. There was no motion or ruling about SCOsource licenses.
Describes witnesses: decades of experience, consistent testimony about value.
Novell witnesses: an in-house lawyer with no engineering involvement. Used the term "forfeiture."
1. SVRX was licensed at no additional fee.
2. No commercial sales of older versions
3. No commercial demand for older versions.
Novell offers no evidence to the value of the older SVRX.
Copyrights are separate. They [Novell] only have entitlement to royalties on certain SVRX licenses.
SCO was unsuccessful largely due to Linux, free, "using the technology" [of SCO] (I thought one couldn't argue at closing things not in evidence. But they managed to get that in there. Sticking with it to the last!)
In plain language, these were UnixWare licenses [not SVRX]
"Trunk" is both old SVRX and UnixWare.
Royalty turns on specific products being licensed.
SCO has the right to distribute UnixWare, which includes SVRX. (UnixWare never reached the threshold where royalties would be due Novell for UnixWare.)
SCO distribution rights include all UnixWare versions. Questions about what's new are not relevant.
One other issue: Novell's effort to shift the burden of proof. He cites 3rd Circuit case (Boeller?): fiduciary dispute, appropriate to shift burden, but this rationale is "out" because Novell was not prejudiced. They had discovery rights; e.g., they could depose Sun and Microsoft.
Novell has the burden. But if SCO had it, we have met it.
1. What part of Sun/MS licenses are SVRX
2. What part of SCOsource were SVRX
3. (I missed it) [PJ: Almost certainly on both lists it would be the question of whether or not SCO had the authority to enter into SCOsource licenses, and specifically Sun's without Novell approval.]
MS Sections 2, 4
(comments about how section 3 was on the list until Novell dropped it)
Sun Section 4
SCOsource division accounting is irrelevant.
Sun/MS deals: UW license including older SVRX that's in UW
Others: license "IP"
MS Section: de minimis
Extend to allow MS to use UnixWare and older SVRX that's included in UW in its products. Also new Open Server license.
MS (& Sun) was only valuing UnixWare, not older versions. [(Was this proved at trial? I missed it.)]
Prior versions were licensed "at no additional charge" (which was the case all the way back to the AT&T days: license cost was the same for one licensee that didn't want older versions as for another that did) (Shows large chart in the courtroom said to be side-by-side comparison of said two licenses. It faced away from me.)
If using prior version for a derivative, a licensee would still pay their royalties according to the modern product license.
Novell never sought an allocation on prior contracts.
Section 2.1: this was a release by SCO (emphasis on "by SCO")
Section 2.2 also limited: license to all owned or licensable by SCO
Sun: de minimis
1994 Sun royalty buy out: 82.5 million to Novell (every cent)
2003 agreement included OpenServer and drivers to enable Sun to do an x86 version. [(I recall that an x86 Solaris was available back in 1996. Maybe he's referring to x86-64 (IA64) or Merced?)]
Solaris: no evidence that non-UnixWare SVRX code is in OpenSolaris. The confidentiality expansion applied to UnixWare. Novell offered no basis for non- de minimis valuation.
SCO promised not to bring claims. Grant from SCO not Novell
If Novell wants to sue Sun, it can. (He did make a comment about IF the copyright ruling stands.)
Contract contains acknowledgment of uncertainty. "Customer is aware that these claims are in litigation."
Basically, they were getting an insurance policy from SCO.
APA also gave Novell's claims related to business to SCO.
One more item on apportionment:
No basis for Novell's apportionment.
(from testimony or deposition): Even when the products were current, they would have cost $700K for Sun, 1.2M MS and another 700K somewhere (I didn't catch it all)
(He shows a chart) These values are when the products were current. This should be a ceiling.
Novell says the broader disclosure rights were not part of this valuation.
True, but the broader disclosure rights were for UnixWare only. [(really?)]
One other source of evidence (for an alternate ceiling):
MS deal section 3: 7 million to SCO (undisputed)
Testimony puts Sun's deal for undisputed SCO material as "at least as valuable"
Thus, we have another ceiling: 3 million = the 10 million Sun deal - the 7 million SCO portion.
SCO's authority to enter into these agreements. They have the authority for "incidental" licensing. Considered dictionary and case law definitions.
One definition: "minor value" part
Instruction from Novell defining "incidental": not SVRX licenses prior to date of APA.
[(But what about a "restatement" of just such a license? SCO position seems to be that restatement is incidental (no additional cost over prior buyout), but this wasn't stated explicitly)]
Incidental according to Novell: "if it's a license of prior versions along with current product, the prior is incidental" (I missed the source of this.)
Novell "kind of" agrees with SCO in this. But: MS/Sun licenses are different from, say, NCR in Novell's view. Not so in SCO's view: there's no difference.
If someone wanted it, a prior version would be provided to a licensee of the current product, there's just no demand.
The 2003 Sun agreement had no effect on the 1994 agreement (which it restates). [(What about broader disclosure?)]
Regarding APA term enabling SCO to do its business with regard to source code, Novell "reads this provision out of the agreement."
Novell is estopped from contesting SCO's authority with respect to SCOsource because it didn't complain back when the deals were made.
Value of non-UnixWare SVRX: witnesses for SCO outweigh those for Novell.
Final: this court's role is not [didn't record the exact term. "In the business"?] of forfeiture.
After the concluding arguments, the entering of exhibits into evidence was handled, including depositions from IBM case. Most were not objected to. The IBM depositions were limited in scope, and SCO didn't object considering the limit. Some exhibits were also identified as confidential, to be filed under seal.
There was one item, #172, not admitted. I missed the objection/reason.
The usual "I am not a lawyer" and "my notes are not verbatim" disclaimers apply.
So Paul was correct. It was Eric Acker who did the closing statement for Novell. Chris says Paul hit all the bases, so instead of repeating what he wrote, he went to the trouble to make a graphic for us of the Venn diagram that SCO used in its testimony on this day and which Chris carefully copied as he listened. He says this: "
In the first diagram, the Solaris circle and arrow leading to it are not shown. In the second, it's there.":