Advertisement
Front, Kernel, Security, Distributions, Development. See your byline here on LWN.net.
Weekly Edition
Return to the Announcements page
| |||||||||||
|
| |||||||||||
Decatur Jones' SCO Analysis (Groklaw)
Groklaw has come up with a copy of a research report on SCO prepared by Dion Cornett at Decatur Jones. It's available in PDF format. This report is very much worth a read; this analyst has at least one eye open. "Furthermore, other courts have ruled that software interfaces are not protected in that their 'fair use' allows for interoperability. Finally, the fact that these header files contain definitions but not functional code sheds more doubt on SCO claims, in our view."
(Log in to post comments)
Huh, doesn't sound too good for SCO Posted Dec 25, 2003 0:12 UTC (Thu) by bojan (subscriber, #14302) [Link] I don't know if it's just me, but this whole report doesn't sound all that good for SCO. Who would have thought that folks like this would take into account what Linux community thinks on the whole matter.But the best part for me was the bit about Novell's claims to copyright. Seriously, how are we all supposed to know who really owns the damn thing (in the imaginary world in which there is actually Unix code in Linux)? Pretty powerful argument when you get sued for copyright infrigement. Darl's push for world domination is slowly melting away...
Decatur Jones' SCO Analysis (Groklaw) Posted Dec 25, 2003 1:44 UTC (Thu) by chill633 (guest, #16013) [Link] Off topic, but interesting...On Dec. 23 the per share value of RHAT passed SCOX. May the trend continue.
An analysis of SCO's claims Posted Dec 25, 2003 1:45 UTC (Thu) by larryblunk (guest, #12587) [Link] Most of the ABI definitions mentioned by SCO originated in Seventh edition Unix or earlier. Some of the defines (such as those in errno.h and signal.h) had their origin in Fourth Edition Unix. There has never been a registration of Fourth Edition Unix with the copyright office. Without registration, Fourth Edition would be considered public domain as, prior to 1978, registration was required in order to received copyright protection. Fourth Edition Unix was released in November, 1973.Further, while the Fifth, Sixth, and Seventh Editions of Unix were registered, the registrations were not submitted until 1992 (by USL). Copyright law requires that a registration by submitted within 5 years of publication in order to receive legal protection. Seventh Edition may have been eligible to receive copyright protection without registration as it was published in 1979. However, AT&T neglected to include Copyright notices in the files. For publications between 1978 and 1989, a copyright notice was required in order to receive automatic copyright protection (without registration). After 1989, copyright notices were no longer required. Here are links to the old Unix code for a number of files mentioned by SCO: The original errno.h defines can be found in user.h from Fourth edition Unix. For those wondering about the comments in the errno.h file, these originated from errlst.c, which also appeared in Fourth Edition Unix (for use with the perror() routine). The original signal.h defines can first be found in param.h from Fourth Edition Unix. A number of ioctl.h and ioctls.h defines are found in tty.h from Seventh Edition Unix. Many of the stat.h defines can be found in stat.h from Seventh Edition Unix. Likewise, you can see the origins of ctype.h in ctype.h and ctype.c in ctype.c from Seventh Edition Unix. By the way, these are part of the first stardard I/O library and have nothing to do with Unix ABI's. Finally, a.out.h is available in in a.out.h from Seventh Edition Unix. Not that Linux now uses ELF format rather than a.out. acct.h is from the Berkeley accounting code. I can't see how SCO could claim this fall under their copyrights. The only questionable file is ipc.h which is apparently derived from the Sys V IPC support.
An analysis of SCO's claims Posted Dec 26, 2003 12:17 UTC (Fri) by gdt (subscriber, #6284) [Link] Hi Larry, My fuzzy memory is that AT&T claimed the early UNIX code was a trade secret and had contract conditions to keep it that way. That would explain the lack of copyright registration (as claiming both copyright and trade secret would be contradictory). Cheers, Glen
The header files claim resembles the IBM PCM dispute Posted Dec 25, 2003 8:33 UTC (Thu) by chel (guest, #11544) [Link] Long time ago there was a dispute between IBM and suppliers of plug compattible mainframes. I don't recall all the details, but I think the effect was that there are limitations on the protection of interfaces.
IBM's mainframe-interface situation applied only to IBM Posted Jan 2, 2004 19:00 UTC (Fri) by Max.Hyre (subscriber, #1054) [Link] Some years ago I worked for a company building equipment to plug into IBM big iron. I'm pretty sure this description fairly represents what happened. We got complete specs for IBM's HW interfaces, for any device or CPU we wanted. These were available thanks to the situation you refer to: when Amdahl, Hitachi, et al., started building mainframes which could directly replace IBM's, operating directly with IBM-built peripherals and protocols, IBM fought back by continually changing the specs, and keeping them trade secret. Unfortunately for IBM, they'd pretty well wiped out the competition (Univac, Burroughs, RCA, Honeywell, etc.), and the upstart Plug-Compatible Manufacturers (PCMs: Amdahl, etc.) brought the Feds' attention to the fact that IBM had a monopoly in the market, and was using it to compete unfairly. Part of the consent decree was that for some time (40 yrs?) IBM was obliged to produce the specs and make them available to all comers. The Feds of that time being rather more serious about the remedy than those of late, we had access to complete information. Note, however, that this applied only to IBM; other [non-monopoly] manufacturers were still free to build funky interfaces and make the competition do the reverse engineering. Thus, IBM's situation doesn't apply to the SCO dust-up, since the only point that's clear is that SCO lacks a monopoly in the market, however much they wish it were otherwise. On a faintly-related note, having the docs still left you far from your design: they were so dense, with so little redundancy, that we'd find ourselves arguing over subtle shades of meaning of a single word, and how the interpretation would affect our design. We came to call sessions analyzing the docs our ``Bible-study classes''. With that background, you can understand why the joke about being given information which is ``complete, accurate, and of no help whatsoever'', lately applied to Microsoft, originally pointed to IBM.
Decatur Jones' SCO Analysis (Groklaw) Posted Dec 27, 2003 0:53 UTC (Sat) by josh_stern (guest, #4868) [Link] Their analysis fails to mention the likely possibilitythat some enterprising law firms will organize all recipients of SCO license demand letters into a large class action suit against SCO. Courts, usually on the side of reducing case load, will go for it and consolidate the cases, and SCO will just wind up on the hook for the defendants' legal expenses.
Decatur Jones' SCO Analysis (Groklaw) Posted Dec 31, 2003 1:26 UTC (Wed) by Pani (guest, #18330) [Link] The SCO CEO Mcbride caused a cluster mess when he was at Pointserve, where he canned the existing business plan, then paid $$$$$ to outside consultants for a bogus business plan that made money off of "internet advertizing" yes pop-ups. (completly new direction for Pointserve, no way to make money either) I talked to him for a hour when I worked at Pointserve, he is All Smoke, Mirrors, and buzz words. He was eventually canned and Pointserve programmers left demoralized. Sorry SCO.
McBride's role Posted Jan 8, 2004 3:57 UTC (Thu) by ccyoung (subscriber, #16340) [Link] but McBride isn't about SCO - he's about Canopy Group
|
|||||||||||