BOINC: lib/cal.h license issue agree with the DFSG?

[Sean Kellogg]

On Saturday 02 January 2010 10:15:19 am Francesco Poli wrote:
> On Fri, 1 Jan 2010 15:13:58 -0800 Sean Kellogg wrote:
> 
> > On Friday 01 January 2010 2:57:18 pm Francesco Poli wrote:
> > > > /* ============================================================
> > > > 
> > > > Copyright (c) 2007 Advanced Micro Devices, Inc.  All rights reserved.
> > > > 
> > > > Redistribution and use of this material is permitted under the following
> > > > conditions:
> > > 
> > > I cannot find any permission to modify or distribute modified versions
> > > of the file.
> > > This seems to fail DFSG#3.
> > 
> > What?! The grant is /right/ there... "Redistribution and use of this
> > material is permitted" provided the following criteria are met, and
> > then it lists the criteria. I suppose it could be its own little bullet
> > point, but that sure seems explicit to me.
> 
> As has already been pointed out by Steve Langasek, redistribution and
> use does not clearly cover modification and redistribution of modified
> versions, which is what I was talking about.
> 
> > That you failed to see that as a grant really calls into question the
> > neutrality of the rest of your license evaluation.
> 
> Neutrality?  We are not on Wikipedia, here!
> I clearly stated that I was going to express my own personal opinion...

Which is exactly why d-l still has a bad reputation in Debian...

> [...]
> > > This takes away a right I would have in the absence of any license.
> > > That is to say, in order to get the permission to redistribute or use,
> > > I must surrender my right to commence or participate in any legal
> > > action related to this work.
> > > I see this as a fee required for getting the permission to
> > > redistribute: the presence of such a fee makes the work fail DFSG#1.
> > 
> > The GPL takes away all sorts of rights... this can't possible be what
> > DFSG #1 is intended on prohibiting.
> 
> Which rights (that I would have in the absence of any license) does the
> GPL take away?

GPLv3 sections 15 and 16. The presence of such waivers are often seen by courts as a form of consideration, which implies it is either a fee or a forbearance. Since it's not a fee in this case, it must be a forbearance, which is the taking away of a right. It's the common position among law professors where I attended that the GPL is a contract, as opposed to a straight up license, because of these very clauses.

> [...]
> > > This clause, instead, seems to say that, if any limitation of liability
> > > is unenforceable, then, boom!, the whole grant of permission is void.
> > > This could discriminate against people living in jurisdictions where
> > > local law forbids too extreme limitations of liability.
> > > If this is the case, then it fails DFSG#5.
> > 
> > This continues to be a laughable argument. The GPL discriminates against
> > countries who jail anyone who uses software licensed under the GPL.
> > Is that discrimination?
> 
> I don't think so: in your example, that law is *specifically* designed
> to attack the GPL, whatever the GPL text may say.  As a consequence, the
> discrimination is not caused by the GPL, but by the law-makers.
> 
> [...]
> > Is this clause really any different than "you aren't allowed to do
> > anything illegal with this software?"
> 
> Steve Langasek has already explained that such a clause is equally
> non-free.

While looking up the specific clauses for disclaimer and liability, I noticed section 12 of GPLv3. Curious as to how that clause is not essentially the same as the non-export clause? As a resident of the United States, I am bound by its laws. As I read (s)12, if those laws prohibited me from complying with a clause of the GPL, I lose the license granted by the GPL. Sure sounds like a "don't do anything illegal" clause to me.

> > > [...]
> > > > This license forms the entire agreement regarding the subject matter
> > > > hereof and
> > > > supersedes all proposals and prior discussions and writings between the
> > > > parties
> > > > with respect thereto.
> > > 
> > > It really seems that no other grant of permission may be considered
> > > valid...
> > 
> > How do you reach that conclusion?!
> 
> Since "this license [...] supersedes all proposals and prior
> discussions and writings [...]", it seems that I cannot consider any
> other *prior* grant of permission as valid.
> 
> Maybe *later* grants of permission can be valid and I should have been
> less fast in generalizing my sentence to *any* other grant...

Ah, perhaps I mistakenly read your comment to suggest there was /no/ grant of permission, as in this clause negated the grant at the top. But, yes, this is a very common contractual construction, since oral agreements are (a) easy to make and (b) binding, this sort of "four corners of the agreement" clause that explicitly excludes any previous negotiations is to be expected in nearly any contract. What's more interesting is it doesn't go on to prohibit modification of the agreement without consent of identified agents.

I guess my question is, do you feel it in someway presents a DFSG problem?
 
> If this is actually your objection, then point taken.
> 
> > 
> > > > This license does not affect any ownership,
> > > > rights, title,
> > > > or interest in, or relating to, this material.
> > > 
> > > I think that this means that, if I hold some right (e.g.: copyright)
> > > on a part of the work, then I am not constrained by the license, for
> > > that part of the work.
> > > This seems to be superfluous to say.
> > 
> > The license is just being very clear that the license in now way
> > diminishes the ownership rights of AMD in the underlying code. Hardly
> > superfluous if you are AMD.
> 
> I think it is superfluous, since no part of the license seems to do
> things like transfers of copyright ownerships or such.
> Hence, it looks like a clause that makes it very clear what is already
> rather clear: in this sense, I think it is superfluous.

Then perhaps you did not go to law school, where they drill into your head that words are cheep, but litigation is expensive. Always better to err on the side over verbosity if there is ever a doubt... and given the FUD that circulates regarding FOSS licenses "stealing" code, I can't say I'm surprised a corporate lawyer would add this extra clarification.

> I may be wrong, of course.
> The clause seems to be harmless, anyway.

-- 
Sean Kellogg
e: skellogg at probonogeek.org
w: http://blog.probonogeek.org