GPL compliance

[Gary W. Swearingen]

Daniel Carrera <daniel.carrera at zmsl.com> writes:

> On Thu, 2006-29-06 at 09:48 -0700, Gary W. Swearingen wrote:
>> You'd better spend a few hours reading and thinking about copyright
>> law.  (There are plenty of intros on the Internet.)
>
> I've already spent a few years. I'm not exactly new to the GPL (but
> IANAL). The first time I read it was 8 years ago, and I read it very
> carefully then and I've spent a fair share of time since discussing the
> legal implications of it, so don't treat me like I'm new to the GPL.

Sorry, but you seem to not understand the most basic priciples of
copyright and licensing.  Maybe we can blame it on careless writing.
But I'll match your suggestion that I read GPL further by recommending
a reading of some of the copyright law.  It's remarkably clear, even
providing some definitions of terms. Here's a couple of links:

http://www.loc.gov/copyright/title17/            (17 USC)
http://lcweb.loc.gov/copyright/circs/index.html  (intro to copyrights)

>> You're supposed
>> to have licenses from of _ALL_ owners of IP on the CD,
>
> If I get two GPL programs, I can give them to you under the GPL as a
> package, and you don't need two copies of the GPL for them. Ubuntu comes
> with 50-100 thousand packages and the packages have at easily a million
> IP owners. You don't need 100,000 or a million copies of the GPL. That

Which is all irrelevant.  It should have been obvious that I was not
talking about COPYRIGHT files or paper documents; I was talking the
many contractual relationships between you and each owner of the IP
from which you intend to copy or derive.  They all have terms to be
followed alike.

> would make any redistribution prohibitive. Suppose I take 2 GPL products
> and put them together as one program. Do you now need two licenses for
> one product?

Depends on what you mean by "license".  If you're talking about copies
of the text of the license, then you're raising an issue for a
different thread. I haven't given that much thought, but I'd guess
that the GPL would be satisfied by a single copy of some licensing
text that says "all code licensed under FSF's GPL v.2", but most
people would at least have some text like that for each chunk of code
(which would be multiple copies of the first part of the license
language, with the FSF's part by reference.  Regardless, you still
have thousands of licenses obey.  You very likely even have licenses
to follow for which the text has been removed even by reference.  (I'm
thinking of Unix and/or BSD code, but I don't have proof at hand.)

Similarly for copyright notices, but few would be bold enough to
remove those, regardless of what any license requires.

>> But if
>> they're nasty licenses with magic statement clauses, then you should
>> (?) worry about your magic statements for all.
>
> That's why I use a distribution that includes only free software.

You say that as if using free software does absolves you of the legal
requirements to obey licensing terms regardless of their
inconvenience.  Of course, you'll be in good company if you take the
risk of ignoring some requirements when it seems very unlikely that
you'll be punished.

I note a clear omission of any argument from you that allows you to
ignore the requirements of _ALL_ of the IP owners.  Do you think you
do not need Red Hat's permission to use their parts of Linux in your
Ubuntu OS according to each and every term of the GPL, including the
source requirements which include the magic statement option?

> And
> these are not "magic" statements. This is section 3 of the GPL.

They're no less magic than the last bytes of an MBR or the first bytes
of an executable.  This is Unix, more or less -- give me a break.

>> To repeat for clarity: That license that covers the CD is only between
>> you and Canonical and only applies to the IP owned by Canonical.
>
> That's not what the GPL says. I recommend that you spend a few hours
> reading the GPL carefully. Pay special attention to section 3.

I agree that the GPL doesn't say it.  Licensing law says it, and I
thing you'll agree if you understand what I'm trying to say.  It is,
of course, true that many licenses cover the CD; that was the basis of
my message to you.  In the quote just above I was refering to the
single license that you seemed to be worrying about, in unjustified
priority to all the rest.  Again, the law considers you contractually
(or at least license-ly) related to each and every owner, by the terms
of their licenses.

> 3(a) is not relevant because Canonical is not relying on it for its CD
> distribution and nor am I.

Again, that seems unlikely, and you've again given me no reason to
believe it.  Be sure to let us know when where you've found their 3b
"written offer", if you find it.


BTW, I'm not opining about whether a GPL-based license of, say, Red
Hat, is satisfied by finding a single publisher like Canonical (who
derives from Red Hat's IP) that offers copies which use the IP.  I was
just trying to convince you that it's something you can worry about as
much as the single license on the whole CD.  (But I suspect that GPL
licensors who interprete it the bad way could be found, especially if
the issue was ever thought to matter much.)