Not logged in
Log in now
Create an account
Subscribe to LWN
Pencil, Pencil, and Pencil
Dividing the Linux desktop
LWN.net Weekly Edition for June 13, 2013
A report from pgCon 2013
Little things that matter in language design
I don't think the GPL can affect code that is not derived. That's all that copyright law allows it to do.
Or am I missing something?
SamyGO: replacing television firmware
Posted Nov 16, 2009 22:38 UTC (Mon) by dwmw2 (subscriber, #2063)
It can grant you permission with conditions attached. Copyright law
does allow that. And unless you abide by those conditions, you don't have permission
to distribute the original GPL'd work.
So I can release code under a licence which says "you can only use this if you send
me a postcard," or "if you name you first-born son after me," or "if all
software you ever write is released under the GPL," or "if I like your
It's my prerogative to grant or deny permission to use my own code as I
see fit, under copyright law.
And I most certainly can release it under a licence which says that you
may only redistribute it if you don't do so in combination with other
software which is under
a different licence. Copyright law most definitely does allow that, and that's what
the GPL does.
Posted Nov 16, 2009 22:48 UTC (Mon) by jake (editor, #205)
Interesting. If I am understanding you correctly, you are saying that anyone who distributes a Linux kernel with a binary driver (many, if not most, embedded Linux devices out there would qualify) is in violation of the GPL.
Not an interpretation I had heard (or, perhaps more likely, understood) before.
Posted Nov 16, 2009 23:09 UTC (Mon) by dwmw2 (subscriber, #2063)
"Interesting. If I am understanding you correctly, you are saying that anyone who distributes a Linux kernel with a binary driver (many, if not most, embedded Linux devices out there would qualify) is in violation of the GPL."
One of the tenuous counter-arguments which is often presented (other than the "copyright law doesn't allow that" fallacy) is the final paragraph of §2, which states that "mere aggregation ... on a volume of a storage or distribution medium" doesn't count.
Some people argue that that "exception" is so wide-ranging that it might as well be rephrased as "Oh, actually just disregard the previous two paragraphs, even where we explicitly spelled out that we meant this to apply even to works which are independent and separate works in themselves; we didn't really mean that."
Some people go so far as to claim that all forms of "aggregation" are permitted by the exception — so since linking to non-GPL'd work is merely a form of aggregation, that's allowed too. Under that interpretation, the GPL would actually become more like the LGPL.
Personally, I feel it's much more likely that that "exception" clause applies to stuff like magazine cover CDs, where mostly unrelated software is just aggregated together for distribution. Or where it happens to sit on the same hard drive or backup tape as if by coincidence.
I definitely don't believe that any sane interpretation of the "mere aggregation" exception can apply to a coherent product where both the GPL'd work and the non-GPL'd work are fundamental and necessary parts without which it could not operate.
But of course nobody is right or wrong until it's been heard in court.
OT - Rant against case law
Posted Nov 17, 2009 12:10 UTC (Tue) by Felix.Braun (subscriber, #3032)
But of course nobody is right or wrong until it's been heard in court.
<completly off-topic rant>
Just for the record, the quoted statement would only be true if courts could never be wrong. That is, if they had an authoritative say on what the law should be.
As a (non common law) lawyer, I take exception at this view. Judges are just opinionated people who have been handed some robes (and in the UK maybe some wigs) by the politicians. They are just as likely to make mistakes as the next guy. They are not inherently more likely to be right than anybody else, even *gasp* non-lawyers. Of course, judges themselves know this, and they therefore appreciate a well reasoned argument, no matter who makes it. Generally speaking, even after having decided a case.
Of course, in practice the legal traditions work out to be remarkably similar. But I really do blame the case law system with it's principle of stare decisis for instilling such an attitude of unreflected acceptance in otherwise critically thinking legal laymen.
SCNR </completly off-topic rant>
Posted Nov 17, 2009 12:47 UTC (Tue) by dwmw2 (subscriber, #2063)
But nobody is right or wrong; it's not black and white. It's just a document written in English and there can be many different interpretations of it; ranging from the sane interpretations to the utterly nonsensical ones such as the "counter-arguments" I mentioned above.
There are very few interpretations which actually matter, though. The first one which matters to you might be the opinion of your company lawyer, when you consider shipping a product which violates the GPL by including both a kernel and a binary-only module in the same coherent whole. Or, as the GPL phrases it, in a "derivative or collective work based on" Linux.
If he is doing his job properly, he will prevent you from doing that. On the other hand if he doesn't do his job properly, you may well end up bumping up against an opinion which takes precedence over his — the opinion expressed by the (final appeal) court when you are sued for copyright violation.
That opinion, ultimately, is the only one which really matters. It doesn't make it right, and there's plenty of examples of cases where the courts have got things "wrong" in the eyes of many people. But it is the only interpretation which really matters in the end, in practical terms.
Unless, of course, you think the revolution is going to be triggered by a copyright case... :)
Posted Nov 17, 2009 14:51 UTC (Tue) by Felix.Braun (subscriber, #3032)
Posted Nov 18, 2009 4:53 UTC (Wed) by butlerm (subscriber, #13312)
That was SCO's theory regarding JFS. It was first developed for a
proprietary Unix licensed from source code SCO claimed it had the copyright
to, therefore (so they claimed) any version for any other operating system
was also a derivative work of Unix System V (and they were owed billions in
I call that the "derivative work by contagion" theory, and it is one of the
reasons why people shouldn't blindly go around claiming things are
derivative works when there is no rational theory for why that should be
the case. Combination may produce a copyright protected collective work,
but the mere fact of combination certainly does not necessarily infect each
of the original components with derivative status when considered
separately, and more especially when the kernel specific binding code is
removed. That would be like saying if a photo is printed in a magazine,
not only is the photo a derivative work of the magazine, the magazine is a
derivative work of the photo.
Posted Nov 18, 2009 10:27 UTC (Wed) by anselm (subscriber, #2796)
The fun thing about IBM's JFS as available in the Linux kernel was that
the JFS code IBM contributed to Linux didn't in fact have anything to do
the JFS code in AIX (which SCO claimed was a derived work of Unix).
The Linux JFS code is a port of the reimplementation from scratch of JFS
that IBM did for OS/2, so it couldn't have been »infected« with SCO-owned
code because it never actually came in contact with any.
Copyright © 2013, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds