Castle Technology UK Ripping off Kernel Code?

Jonathan Riddell writes "`It would appear that Castle Technology Limited, UK, have taken some of the Linux 2.5 code, and incorporated it into their own product, "RISC OS", which is distributed in binary ROM form built into machines they sell. This code is linked with other proprietary code.' Full details from Russell King on lkml."

it was bound to happen

[stonebeat.org]

if you make any code opensource, you should be prepared for other's to copy it.

Now let us see what GPL does.....

Re:Sue them

[Black Copter Control]

Who would be the initiator of such a lawsuit?

It depends on who holds copyright to the associated pieces of code. Best bet is that it's been assigned to the EFF, but it could also be Linus and/or some of the people who wrote the bulk of the code.

It'll actually be rather interesting (in ~200 years) when it comes time to determine when the code's copyright expires. Just who's lifetime does each piece of code expire in relation to?

Re:it was bound to happen

[Majin Bubu]

Surely they can copy and/or modify it, but if they want to distribute (for free or for a fee, it does not matter) the derived code in binary form, they must release the source as well, otherwise they violate the license.

Re:Does that mean...

[bwt]

No, it means that they are commiting willful copyright infringement for commercial gain. The penalties for that are severe and include the larger of statutory and actual damages. The statutory damages can be up to $100K, iirc. Actuals include any revenue which results from the infringement.

I hope somebody tears them a new sphincter, if this is true.

Re:For damages?

[Beryllium Sphere(tm)]

Copyright law allows for something called "statutory damages", which means that if someone infringes on your (registered) copyright you can collect a fixed amount without having to document monetary loss.

"Nobody lost anything", except control over their creations.

Copyright grants a partial legal monopoly on distribution of the copyrighted work. The owner can make people pay for it (the usual approach), or make them accept the GPL, or even prevent circulation of the work altogether (the way Sinatra pulled the movie "Suddenly" off the market after the Kennedy assassination).

Typo - FSF, not GNU

[wass]

See subject.
Tis the folks at FSF that challenge GPL violations. Of course not folks at GNU, since GNU ain't no place or organization.

Re:Sue them

[mewsenews]

As another reply said, it's the FSF not GNU, and I've also heard that Stallman will not take on a legal case involving the GPL unless the copyright for the GPL software has been signed over to the FSF. I might be wrong.

Re:How to prove anything?

[bwt]

During the discovery phase of the trial, the defendent would have to produce the complete source code and build instructions for their product. The plaintiff would have an expert follow the build instructions and verify that they result in the exact exectuable that the defendent ships. Then the expert would examine the source code for "substantial similarity" to the copyrightable elements of the linux kernel code. A judge would hear this testimony and rebuttals and examine the evidence it was based on.

Legal arguments on affirmative defences of fair use and licence compliance could be made. The judge would rule on infringement, then if the plaintiff prevails, he would rule on damages. Factors influencing damages would be willfulness of the infringement and the presense or absense of commercial gain as a result of the infringement.

Re:Does that mean...

[ubernostrum]

Actually that's exactly how it works. For better or for worse.

Nope. Consider the physical-property equivalent: I steal something from you. Do you have the right to come steal it back? No. Similarly, if they "steal" GPL'd code, we don't have a right to steal it back; the legally correct course of action is to file suit, get an injunction against distribution of the stolen code, and recoup losses from the infringer.

Re:Does that mean...

[bwt]

You are confusing the conditions for complying with the licence with the penalties for infringing the copyright.

The judge *could* order compliance with the licence as part of the penalty phase, but it is much more likely that he would award monitary damages.

The interesting thing is that each patch to the linux kernel could be viewed as separately copyrighted by whoever the author of that particular patch is. The statutory damages can reach $100K per violation if the judge wants it to.

Re:GPL Question

[Corydon76]

Any program that uses the GPL'd code HAS to be available for free in both binary and source form?

No, if the binary is distributed for a fee, then the source must be available, even if it's for a nominal fee. Charging you for the media and shipping is perfectly acceptable. However, charging you $40 for the binaries and $200 for the source is NOT acceptable, as the source fee is clearly not reasonable, compared to the charge for the binaries.

Charging for GPL code is not forbidden. You only need to provide the source.

Re:Who files a lawsuit?

[On Lawn]

As I understand it that couldn't be further from the truth. Linus has intentionally let everyone keep copyright of their pieces. He said this was to keep one company from buying him (or any two or three) developers.

That makes him different then SleepyCat and others who retain copyright so they can release code under any copyright they wish to (which I also support as useful in some situations).

-------------------
OnRoad [onlawn.net]: The automotive magazine for before your ride home.

Re:Who files a lawsuit?

[bwt]

Absent an explicit, signed contract transferring copyright ownership, the author retains copyright for those elements of the work which are his original expression. *Use* requires a licence, ownership transfer requires a signature.

Some people do sign over their copyright ownership when they participate in GPL projects, but more often people don't. The crux of the issue is whether you can *change* the licence. For example, the FSF can update the GPL and the changes take effect immediately on any works they own the rights to.

Assuming that no explicit transfer of copyright ownership has happened, the authors of the particular code that was copied probably each have an independent cause of action against the plaintiffs. That could hurt the plaintiffs, because the statutory damages can be calculated for each act of infringement.

Re:Some may argue...

[(void*)]

The problem with that argument is that it betrays a lack of understanding of the rewards of free-software. When people share software, they expect IMPROVEMENTS. The software is GIVEN AWAY, so that improvements, in the form of making the code more efficient, more portable, more usable, are all the "payment-in-kind" for software. What is damaged is not that there's erosion of the user base (there is) but it is also that one can have any expectations. That someone can use something to imporve himself alone, and forget about the rest of the world.

Re:Castle Technologies?

[paulc0001]

Wouldn't the company in violation be RISCOS Ltd? [riscos.com]

Actually, no. RISCOS Ltd are responsible for RISC OS 4. The Iyonix PC uses RISC OS 5 that was developed by Castle Technology (well, developed for them but not by RISCOS Ltd).

No... read...

[mindstrm]

No.. the GPL ONLY gets it's strength from copyright law. IF copyright law allows it, the GPL can be IGNORED.

The GPL is the set of terms under which you can do things OTHER than what you are allowed to do under standard copyright.

It is not a use license, you don't agree to it in order to use said software; it is a license that grants you extra rights beyond what copyright does should you CHOOSE to use it (abide by it's terms).

Re:Some may argue...

[jedidiah]

Then this falls back onto straight copyright, in which case this company needs permission from EVERYONE that ever contributed to that part of the kernel. The work in question is not Public Domain. It is still copyrighted. That by itself carries restrictions that make corps like this liable for damages.

This is something that could also concievably involve criminal charges.

Your argument SUCKS

[unicorn]

The law doesn't say anywhere that you're allowed to acquire copyrighted material without paying for it, as long as it's not for commercial purposes.

The law says that it's a violation of copyright law to participate (as sender or recipient) in copying the property of another entity.

Re:Who files a lawsuit?

[necromaedian]

I'm no expert on the GPL but this sounds correct to me. Which is probably why Mr. King is asking for "[any] peoplewhohavetouchedanyofthefollowingcode:
-PCIsubsystem
-IOresourceallocation",
to come forward.

Re:They have nerve

[Paul Vigay]

Actually, speaking as the webmaster of www.riscos.org, that quote is mine, and I have no link to Castle Technology at all. If you're going to call someone names, at least check your facts to see who you're accusing.
www.riscos.org is completely independent and in no way affiliated to ANY RISC OS hardware manufacturers.

Re:Who files a lawsuit?

[Bruce Perens]

Any one of them or a number of them collectively may file suit. All cases have been settled out of court, we've always won - meaning the other party has always removed the infringement, or applied correct terms to their own code.

Bruce

Re:Who files a lawsuit?

[Bruce Perens]

No, Linus has the Linux trademark. He only has the copyright for his own code.

Re:Who files a lawsuit?

[Bruce Perens]

Any one of the copyright holders can sue, or a number of them can band together and sue. We've always won, but the cases have never gotten to court. Their side's attorney concludes that they'd lose the case, and they settle before going to court. They either remove the infringement and apologize, or they place the linked code under a GPL-compatible license. Note that this need not be the GPL, the X license and the LGPL are GPL-compatible, so are a number of other licenses.

Bruce

Re:Sue them

[prizog]

AFAIK, no kernel code has been assigned to the EFF. But you really meant the FSF, which has most of the S390 stuff, but nothing else.

But if you're a kernel hacker, especially in the core, and want to see the GPL get enforced more effectively, just write to [e-mail], and assign copyright to the FSF. Right now, almost everyone who uses the kernel also uses various GNU userspace applications. That's what lets me do my job of enforcing the GPL. But it would be much easier if the FSF simply had copyright in one or two core kernel files.

A sale of a copy

[yerricde]

You don't own the SW after you buy it.

Then what about the Adobe v. Softman precedent, stating that if it looks like a sale of a copy and quacks like a sale of a copy, it's a sale of a copy?

A consumer of mass-market software surely doesn't own the copyright on the program, but unless there's a specific rental agreement between the consumer and the copyright owner, the consumer does own a copy of the program. (A "copy" is defined as the medium in which the program is fixed.) And if the consumer owns a copy, then the defenses in 17 USC 109 [cornell.edu] and 17 USC 117 [cornell.edu] become available, but the right to distribute modified versions is not among them.

Re:Castle Technologies?

[Gerph]

The sources used by RISCOS Ltd are not those used by Castle; the sources that RISCOS Ltd use were licensed a number of years ago from Acorn and do not include any components which are GPL, and acknowledge the components which they do use. Castle have, alledgedly, licensed their sources from Pace Microtechnology plc (the company who bought Acorn) and presumably added the PCI code from the Linux kernel themselves. I say presumably because I can't say one way or another whether the work was done by Pace or by Castle. Castle, as publishers of the product, are those who must be looked to for answers.

Re:What is the oldest piece of code in the kernel?

[Black Copter Control]

The original code (or work) would still be copyright it's original copyright date. The later code would be copyright it's new release date. If you can disambiguate the two (for example, by finding a copy of the original), you get the advantages of the original copyright date.

More info

[Anonymous Coward]

From their web page
http://www.iyonix.com/32bit/opportunities.sh tml

...

Most developers are now porting their software to the IYONIX pc and charging a small upgrade fee per product as well as considering new applications and some suggestions for USB, PCI and software products are shown below.

....

Third parties are encouraged to develop drivers and applications to use PCI cards. In the past it was sometimes difficult to justify developing specialist expansion cards for a relatively small market, but the existence of a wide range of low-cost PCI cards provide a ready solution and the source code for many device drivers is already freely available thanks to Linux.

No mention of open source software or GPL, just freely available.

Re:why violating open source copyrights SUCK

[Anonymous Coward]

Internal use isn't kosher as a matter of opinion, it's indisputably allowed by the license (due to the privacy issues stemming from being required to distribute private work).

Re:It's not though

[Archie Steel]

I think you should read that GPL again. It does not give someone the right to distribute the code (or parts of the code) in binary form, but then requires whomever distributes it to also offer the sources as well - and give credit to the original authors. They are also prohibited from not extending these same distribution rights and obligations to who get their software. So that company can use parts from the kernel, but they have to give out the sources of their software and give credit where it's due - as well as the permission to redistribute the software under the GPL. If I understand correctly, they have failed to do this.

In other words, yes, you were wrong, and have now been corrected. GPL does not mean Public Domain (even though the two share some similarities). It seems you're going to keep waiting a looong time for whatever gripe you have with the GPL to materialize.

A Brief History of RISC OS

[RobinWatts]

A brief history lesson, that might help to make one or two things clearer. Its as accurate as I can manage, but don't sue me for slipups, right? Years ago there was Acorn Computers Ltd; they produced the BBC Micro, BBC Master etc. They then set about designing their own 32 bit cpu for their next generation of computers, and the ARM chip was born. ARM Limited was formed as a spin off and carries on administrating the architecture today. Acorn went on to use the ARM chip in various computers; its first OS was called Arthur, which later developed into RISC OS. Acorn used RISC OS both in desktop targetted machines and increasingly in a range of set top box like products and other embedded devices. The first bunch of ARM chips developed ran in what was called '26 bit mode' (don't worry about it), and the OS largely depended on some custom support chips. Eventually Acorn took the decision to drop out of the desktop market; when it did so, it granted a license to a new company (RISC OS Ltd) to the latest version (4) of its software. RISC OS Ltd carried on developing this OS and continue to do so today as part of their Select scheme. Shortly after granting the license, Acorn split into 2 sections; one became element 14, and the other was taken over by Pace. Pace carried on developing RISC OS in house, and produced a 32 bit hardware independent version of RISC OS. Later, AIUI, Pace stopped RISC OS development work, and shortly afterwards Castle announced that it was bringing desktop machines to market using the Pace version of RISC OS. Some sort of license deal had been made between Pace and Castle - my understanding is fuzzy here, I don't know if its a direct license or whether Castle have a license through another third party. It is this latest version of RISC OS that the accusation of use of GPL'd code has been made against. RISC OS Ltd, ARM Ltd, and Element-14 (AIUI now subsumed by Broadcom) are *not* implicated in any way - don't send them hate mail! riscos.org are *not* implicated in any way - thats a web site run by enthusiasts. Its not clear to me at which point the GPL'd code got introduced into the source tree; it could have been at any point after the RISC OS Ltd tree was forked from Acorns original source. Also, for those that don't know, Russell King was one of the prime instigators of ARM Linux, so together with Linus himself is probably one of the most significant opinions on this matter. FWIW: I'm an independent software contractor, who has worked under contract for Acorn, Pace, e-14, and RISC OS Ltd (though not on the portion of the code under discussion). I have no axe to grind here - this post is purely intended to sort out who the players in this particular drama are (and aren't!)

RISC OS is not Castle

[radiac]

The sad thing is that I had really heard nice things about RISCOS... if it were possible to get an old machine inexpensively in the US, I would have tried it.

That's a bit like saying that you would not buy a second hand 486 with MS DOS on, because a different PC manufacturing company ripped off some GPL code last year to get part of their new system working.

I think a lot of people are getting Castle [castle.uk.co] confused with RISC OS [riscos.com], and in this case even Acorn, who don't exist any more - the code being discussed is for the new Castle Iyonix [iyonix.com] machine, released just before Christmas 2002. It is the first 32 bit RISC OS machine, and as such needs a 32 bit OS. Castle have released RISC OS 5, which is based on RISC OS which is licensed by Pace, who bought it from Acorn just before they were closed down (about 5 years ago I believe).

It is RISC OS 5 that has the alleged GPL breach; previous versions of RISC OS have nothing to do with Castle, apart from Castle's machines run RISC OS.

As a general comment, it would be nice if people on slashdot spent a little more time looking into the facts before posting.

On a separate note, if you really are interested in getting an old Acorn machine, there are a lot of second hand Acorn machines available at extremely reasonable prices if you're willing to spend a little time looking for it - for example, newsgroups, community websites, magazines (such as Archive [archivemag.co.uk], which has a small ads section), or companies which sell second hand RISC OS machines, like CJE Micros [cjemicros.co.uk].

HP Printer license

[tundog]

Just installed an HP printer and the following came up with the EULA. I wonder, if it ever turned out that they infringed on Open Source (e.g. GPL), if it would effect the entire license, making it null and void, i.e. letting me disassemble until my hearts content (clause 4).

8. U.S. Government Restricted Rights. The Software has been developed entirely at private expense. It is delivered and licensed, as defined in any applicable DFARS, FARS, or other equivalent federal agency regulation or contract clause, as either "commercial computer software" or "restricted computer software", whichever is applicable. You have only those rights provided for such Software by the applicable clause or regulation or by these License Terms.

Re:It's not though

[Vicegrip]

If the GPL is held invalid, for whatever reason including patent infrigement, then nobody can use the code without going back to the copyright holder and negotiating a new distribution license. The GPL gives you the right to use code, it does not give you any form of ownership on it.

The only work you own, is the work you do yourself.