On with the interviews and discussions

Bill Claybrook, Abderdeen Group, Discussions

MozillaQuest Magazine: In your report, you say:"

"It appears that there are at least two ways that System V licensees (as well as those who have not licensed System V) can risk legal confrontation with SCO:

1. The copying of code from System V into Linux

2. The placing of derivative works (code) of System V into Linux."

Does that mean that GNU/Linux end-users do not risk legal confrontation with SCO, even if there is SCO-owned Unix code in the GNU/Linux distribution they use?

Bill Claybrook: I said "at least" two ways. I would not make the assumption that SCO would not go after the users that you mention. However, SCO has said that they are primarily interested in the enterprises that are reaping financial benefits from what they consider to be illegal use of their code. Good question, but I could only speculate because I don't know how far SCO is willing to go with this.

MozillaQuest Magazine: How about Linux kernel, GNU/Linux OS, and Linux distribution developers?

Bill Claybrook: Same answer.

MozillaQuest Magazine: In your report, you say:

"This analyst was one of three analysts to whom SCO showed the System V source code (and comments) that it alleges was directly copied into Linux by a large IHV (not IBM or Sun). I neither attempted to verify that either stream of code was from Unix or Linux, nor did I try to verify when the code was written, etc. This is for the courts to determine.

I am a former Unix kernel programmer and the code that I was shown was from a well-known Unix .c file. It is only one of several instances that SCO alleges that an IHV directly copied System V source into Linux. Based on what SCO showed me, the amount of alleged copied code and comments in the .c function amounted to about 80 lines. SCO claims that other IHVs have directly copied System V code into Linux." [Please see Comments Note in the sidebar.]

MozillaQuest Magazine: Is the SCO-owned Unix code that allegedly was copied into "Linux" copied into the kernel.org Linux kernel, some other hacked or patched kernel, the GNU/Linux operating system, a Linux distribution, or elsewhere?

Bill Claybrook: The code that I was shown was reputed to be from Unix System V source and from Linux source that had been modified by a large IHV. I did not have the source code for the 2.2, 2.4, and 2.5 (development) Linux kernels with me to compare, nor do I recall more than a few thousand lines by memory (my limit), but SCO said that the Linux code is in those three kernels which may tell you something about how long ago this alleged copying took place.

MozillaQuest Magazine: Were there any IHVs involved in the original Linux kernel development by Linux Torvalds and others in the early 1990's?

Bill Claybrook: I don't know, but my guess and it is only a guess, that there probably was not if you are talking 1991-1993. I have also not heard of any, so likely no. Please see sidebar Note.

MozillaQuest Magazine: Since it is IHVs that have allegedly copied Unix code into the "Linux" code-base, does that mean that the Linux kernel was not derived from Unix as SCO-Caldera claims?

Bill Claybrook: I did not say that only IHVs have allegedly copied Unix System V code into Linux. I don't have a list of any of the alleged violators, except SCO did say that IBM has copied Unix System V code into Linux. I personally find that hard to believe because IBM has one of the best screening processes of any major supplier for making sure that code does not move into Linux. And SCO said no to IBM's copying the first time I asked the question and then several hours later changed their answer to the one that I just gave. (Emphasis added. This is not the first time SCO's Chris Sontag has changed what he says.)

MozillaQuest Magazine: How about the GNU/Linux operating system?

MozillaQuest Magazine: This interviewer is somewhat confused here. You say:

"copied into Linux by a large IHV (not IBM or Sun)."

But later on you say:

"One example of derivative code that SCO says IBM has released to Linux is IBM's AIX journaling file system." (JFS) and " Two other examples of derivative code that SCO says IBM has released to Linux is NUMA (nonuniform memory access) code and RCU (read, copy, and update) code developed by Sequent prior to IBM's acquisition of Sequent in 1999."

It seems that in one instance you are saying IBM is not one of the IHVs that put SCO-owned code into "Linux" and then later it seems that you are saying that IBM allegedly put SCO-owned code into "Linux". Could you explain that please?

Bill Claybrook: I was referring only to the code that I saw. The specific code that I saw was not copied by IBM, according to SCO, but by some other large IHV. Later they did say that IBM had copied source code from System V to Linux, but I saw no examples of that. (Emphasis added.)

MozillaQuest Magazine: If you are referring to an IHV other than IBM, how did IHVs (independent hardware vendors) get SCO-Owned Unix code into "Linux"? Or is this code tainting restricted to products the IHVs sell?

Bill Claybrook: I don't know how they got the code into Linux. If the large IHV did copy the code as SCO says, then, presumably, a developer or developers working for the large IHV was working on Linux and copied the code. I have no knowledge of who the IHV is or what products they sell or if they even sell Linux products (however, I would assume that they do).

MozillaQuest Magazine: You say,

"the code that I was shown was from a well-known Unix .c file",

and that you were shown only 80 lines of code. Are any or all of these 80 lines of code already in the public domain and/or well-known enough so as not to be trade secrets?

Bill Claybrook: I did not say that I was shown only 80 lines of code. I was shown much more code than 80 lines plus comments but about 80 lines of code/comments were identical in the two sets of code. The code that I saw was in a very common Unix kernel function that anyone who has done Unix or Linux kernel development has seen. I don't get the connection with respect to trade secrets. We are just talking about some C code here that SCO claims was copied. (Emphasis added.)

Bill Claybrook: The code and comments that were allegedly copied were not 80 contiguous lines of code and comments. The comments in some cases had another sentence inserted into the allegedly copied comments, but in the sections of the code that were copied, the code was identical with no inserted code.

MozillaQuest Magazine: Are these 80 lines of code that you were shown exactly identical in SCO-owned Unix and "Linux" or just similar?

Bill Claybrook: The lines of code that were claimed to be copied were identical even the spacing was identical.

MozillaQuest Magazine: To the extent that SCO owns the Unix copyrights and/or patents, do these 80 lines of code out of, what, maybe four-million lines of "Linux" code, suffice to constitute either a copyright infringement and/or a patent infringement?

Bill Claybrook: I'm not a lawyer. I don't know how many lines it takes. But SCO claims that there are other pieces of code that has been copied.

MozillaQuest Magazine: To the extent that these 80 lines of code are under SCO-owned Unix copyrights or patents, how easy would it be for the Linux kernel and/or GNU/Linux developers to yank that code and replace it? How long would it take to do that?

Bill Claybrook: I think for this particular function, it would not take long. I couldn't do it anymore, but maybe you could. But for derivative code like the NUMA technology code that SCO claims IBM has moved to Linux, it could take many person years.

MozillaQuest Magazine: This is particularly interesting:

"I specifically asked SCO if they had any evidence that IBM directly copied System V source code into Linux. The reply was no. SCO has subsequently changed that reply to, 'We have that code but we have not presented it at this time'."

MozillaQuest Magazine: Who made those statements? How much time elapsed between the "no" and then the revised "We have that code" statement?

Bill Claybrook: I asked the question of Chris Sontag. Several hours.

Please see the first two parts of our series about SCO-Caldera's IP claims plus its intentions to enforce and license its intellectual property rights.

SCO-Caldera & the GNU/Linux Community: The SCOsource IP Matter

SCO-Caldera & the GNU/Linux Community: Part 2, Under the Iceberg's Tip