LWN: Comments on "Holiday cheer from the SCO Group"

Basics of copyright

bojan — Wed, 07 Jan 2004 23:45:07 +0000

This part seems to be escaping SCO all the time. You cannot copyright numbers, words, letters or anything like that. So, one could simply write:

#define ONE 1
#define TWO 2
#define THREE 3
...

#define EPERM ONE
#define ENOENT TWO
#define ESRCH THREE
...

And avoid copyright infringement altogether. That is providing any such header file can be copyrighted in the first place and Linus actually pinched those files, both of which seems not to be true.

Why do they even bother to come up with this kind of nonsense, I fail to understand.

Holiday cheer from the SCO Group

jrrk — Tue, 06 Jan 2004 08:59:44 +0000

/* error number .H version 2.8.x */

/* how about this version - to be sure to break K&R compilers */

#ifdef __STDC__
enum {ENOERR,EPERM,ENOENT,ESRCH,EINTR, <etc.> };
#else
#error "Be sure to compile with ansi-C for maximum political correctness */
#endif

SCO copyright claim over call transcript

smoogen — Wed, 24 Dec 2003 17:39:50 +0000

They will probably state that because they said this, by saying anything you were accepting their claim on copyright of the call and thus handed over your copyrights.. probably get them laughed out of court, but hey that will be 10 years from now after all the appeals just to get it listened to.

32V basically declared public domain by USL-BSD judge

tseaver — Wed, 24 Dec 2003 14:07:17 +0000

I particularly like this quote from the opinion, which seems eerily like
what Linus had already done:

> There is an enormous difference between an expert
> programmer sitting down with a pile of textbooks and disjointed
> segments of code to write out an operating system from scratch, and
> that same programmer downloading the operating system intact from a
> public network. In the first case, the programmer could expend
> large amounts of time writing, testing, and debugging the
> newly-created system, with an uncertain prospect of immediate
> success. But in the second case, immediate success would be
> virtually assured. Thus, even if all of the pieces of the 32V code
> had been thoroughly revealed in publicly available literature, the
> overall organization of the code might remain a trade secret unless
> it too had been disclosed.

Just a thought about the headers...

flewellyn — Wed, 24 Dec 2003 13:02:27 +0000

The Linux headers COULD be changed so that the error and signal numbers are written in hex, instead of in decimal. :-) This would, after all, break absolutely nothing (a number is a number), but would make the headers LOOK different. And since SCO seems to be claiming that "anything which looks like ours is ours", the kernel developers could do this and say "nyah nyah, doesn't look like yours!"

Yes, I'm joking. :-)

Holiday cheer from the SCO Group

iabervon — Wed, 24 Dec 2003 05:25:02 +0000

Linus says he generated errno.h with a simple C program, so the file
itself is his work. The comments in the file, however, come from libc
2.2.2's list of strings to implement strerror(). Nobody has yet reported
where J H Lu got those strings, but it probably was from POSIX.

SCO copyright claim over call transcript

gdt — Tue, 23 Dec 2003 11:38:26 +0000

How can SCO claim copyright on the analyst conference call?

Firstly, how does SCO gain copyright of the analysts' questions?

Secondly, how can SCO enforce such a claim without running foul of stock exchange listing rules on concurrent disclosure? (You'll need to forgive the lack of specifics here, since I'm from Australia and each exchange differs in detail. But essentially, the exchange rules require the same disclosures to be made to all stock holders reasonably simultaneously.) A question to the exchange which lists SCOX seems in order.

Oops: Ignore me

Ross — Tue, 23 Dec 2003 07:31:22 +0000

I somehow missed the paragraph with this sentence the first time I read the
article: "SCO will have a hard time convincing a judge anywhere that
copyrights can protect this sort of code".

Assumption

Ross — Tue, 23 Dec 2003 07:07:25 +0000

There is an assumption in this analysis that this list of #defines is
copyrightable. I'm not sure that is a safe assumption to make. How
exactly could another compatible implementation be made that didn't use
the same names and numbers? Would any order other than increasing
numeric value make much sense?

The only remaning possible differences are whitespace (including comments)
and unrelated code. I would hope that errno.h would not contain any
unrelated code. As for whitespace, that was actually different.

IANAL.

What timing!

jre — Tue, 23 Dec 2003 04:10:11 +0000

Wasn't it only yesterday that a LWN reader asked (about the RIAA subpoena ruling): "While I do find this interesting, what does it have to do with free software?"

Here's your answer, straight from the horse's [mouth]:

"DMCA liability extends to those who have reasonable grounds to know that a distribution (or re-distribution as required by the GPL) of the altered code or copyright information will induce, enable, facilitate, or conceal an infringement of any right under the DMCA."

I don't mean to disparage a perfectly innocent comment, only to point out that the DMCA is indeed a dangerous thing to leave lying around. When it falls into the wrong hands, free software likely will be affected.

Holiday cheer from the SCO Group

error27 — Tue, 23 Dec 2003 01:43:03 +0000

If you look at the first version of Linux errno.h had a comment that sort
of says how the file was created.

include/errno.h:
/*
* ok, as I hadn't got any other source of information about
* possible error numbers, I was forced to use the same numbers
* as minix.
* Hopefully these are posix or something. I wouldn't know (and posix
* isn't telling me - they want $$$ for their f***ing standard).
*
* We don't use the _SIGN cludge of minix, so kernel returns must
* see to the sign by themselves.
*
* NOTE! Remember to change strerror() if you change this file!
*/

Holiday cheer from the SCO Group

simlo — Mon, 22 Dec 2003 22:03:19 +0000

I checked the errno.h on three Unix systems I just happened to have accounts on, OSF1 V4.0, SunOS 5.8 and IRIX 6.5.
On the two latter the identation was (not suprisingly) as in SysV5 as stated above and the comments the same to those of Linux's. The comments seems to be identical to the definitions in POSIX though, so there is no proof that the guy doing the Linux version copied from a SysV5 derived Unix and changed the
identation.
In OSF1 the identation was one less than in Linux's but the comments were different.

32V basically declared public domain by USL-BSD judge

freeio — Mon, 22 Dec 2003 20:46:47 +0000

Dennis Ritchie's web site has the following link to the 1994 USL-BSD judgement:

http://cm.bell-labs.com/cm/cs/who/dmr/bsdi/930303.ruling.txt

Note that the judge, in essence, stated that 32V was neither copyrighted, nor could it contain any trade secrets, due to the mishandling of the licensing and labelling by AT&T. As such, anything which is identical to the files in 32V is quite legal to use or copy, without any attribution whatsoever.

From a legal standpoint, SCO has an impossibly high barrier to overcome in the face of the 1994 ruling.

Holiday cheer from the SCO Group

ncm — Mon, 22 Dec 2003 20:15:01 +0000

About the loss... a posting on Groklaw claimed that SCO executive bonuses depended on achieving four consecutive profitable quarters, making this loss a major setback for the individuals involved.

Court cases

jamienk — Mon, 22 Dec 2003 20:13:14 +0000

Will any of the current court cases (against IBM, IBM against SCO, Red Hat against SCO) relevent here? If SCO loses all of them, do they lose on the legal questions raised here? (If losing those cases bankrupts them, that's one thing, but in terms of the legal questions raised...?)