[License-discuss] [Non-DoD Source] Re: U.S. Army Research Laboratory Open Source License (ARL OSL) 0.4.0

Wed Aug 17 13:50:08 UTC 2016

> -----Original Message-----
> From: License-discuss [mailto:license-discuss-bounces at opensource.org] On 
> Behalf Of Engel Nyst
> Sent: Tuesday, August 16, 2016 4:42 PM
> To: license-discuss <license-discuss at opensource.org>
> Subject: Re: [License-discuss] [Non-DoD Source] Re: U.S. Army Research 
> Laboratory Open Source License (ARL OSL) 0.4.0
>
> On Tue, Aug 16, 2016 at 9:43 PM, Karan, Cem F CIV USARMY RDECOM ARL
> (US) <cem.f.karan.civ at mail.mil> wrote:
> > OK, I see where you're coming from now.  I had to have the ARL Legal
> > team explain this to me as well, but the ARL OSL is actually a
> > contract, and the contract can apply even if there is no copyright.
> > We release material to our collaborators on a regular basis under
> > contract; we even do this with software, even though it is in the
> > public domain.  If they break the contract, we can sue them, but we
> > can't sue anyone that they delivered the software to (it's in the
> > public domain, so we don't have any copyright protections to sue
> > over).  The ARL OSL extends this as a chain; the USG releases the
> > software to anyone that wants to download it, but by downloading it,
> > they agree to the contract.  That person in turn can hand off the
> > software to another person, forming the chain.  However, if the chain
> > is broken, the USG only has the right to sue the first person that
> > broke the chain; the others may be able to claim that they got the
> > software in good faith.  Since there is no copyright involved, and
> > since they didn't break the contract, they are innocent; only the
> > person that broke the chain originally is liable (note that I'm not a
> > lawyer, and may have gotten some of this wrong; it's just my understanding 
> > from the ARL Legal team).  This means that to sue, the USG
> will need to prove that the person was the first one in the chain to break 
> the contract.
> >
> > Copyright is something entirely different from contract law.
> > Copyright is a bundle of rights that an author gets by creating a
> > work.  The license allows a user to use the work without getting
> > sued/stopped/etc.  The trick is that since copyright attaches to a
> > work AND since you can't copy/use/display/perform/etc. a work without
> > permission from the copyright holders, you have to be able to point to
> > the license that allows you to use the work without being sued.  That
> > means that a copyright holder doesn't need to follow a chain, it just
> > needs to demonstrate that it has copyright on the work, and that its 
> > license is being violated.
> >
> > The closest analogy I can provide is that contract law is innocent
> > until proven guilty, while copyright is guilty until proven innocent.
>
> I understand the intention, and I know it seems tempting to work via 
> contract, but here's the problem:
>
> Caution-https://www.law.cornell.edu/uscode/text/17/301
>
> "On and after January 1, 1978, all legal or equitable rights that are 
> equivalent to any of the exclusive rights within the general scope of
> copyright as specified by section 106 in works of authorship that are fixed 
> in a tangible medium of expression and come within the subject
> matter of copyright as specified by sections 102 and 103, whether created 
> before or after that date and whether published or
> unpublished, are governed exclusively by this title. Thereafter, no person 
> is entitled to any such right or equivalent right in any such work
> under the common law or statutes of any State."
>
> Some claims of breaches of contract will fall squarely into what this 
> paragraph says: they would claim the same thing as the rights under
> copyright.
>
> In other words: if A tries to make a contract with B, where A says "you 
> can't reproduce this work", that obligation lives or dies via
> copyright alone. (if nothing else is involved)
>
> From what you say, you intend here exactly that: to recreate the rights to 
> reproduce, distribute, or make derivative works, or to put
> obligations as if you had them, through contract. It seems to me that 
> copyright law already says USG can't do that.
>
> You can do a lot of contracts, to be sure; just not those who simulate 
> copyright.

Got it; I'm going to forward your comments to the ARL Lawyer I'm working with 
to see what his opinion is. He's on vacation for another week though, so I 
won't be able to give you a good response until then.

> Caselaw on this exact topic seems a mess. I don't know what would come of 
> this; without getting into it, here's my suggestion, considering
> all I understand from your intentions:
>
> The interesting thing with your intended license/contract is that preemption 
> doesn't matter for malevolent contributors: you can STILL
> make it so that contributors will provide their (presumably copyrightable) 
> work under it, in your projects.
> Because only clauses
> 2 and 4 would be affected by preemption, redrafting the license/contract so 
> that the rest stands in all cases should give you the same
> effect (or close).
>
> Apache license is almost unique in the following respect: there are 2 
> explicit directions in which it works.
>
> Direction (1) - from USG/others to the world.
>
> Here you have the problem that if you start without copyright, and the 
> license tries to usurp the domain of copyright rights, that can make
> it all fail. You said it yourself: the concern is that it depends on 
> copyright, and thus may all be deemed invalid.
> Indeed, I'm just saying that recreating copyright-like rights via contract 
> where title 17 clearly denied them,
> can also be deemed invalid.

Definitely bringing this up with our lawyers!  My intent was that if there was 
copyright, then we could stand on copyright law; if there wasn't, then we'd 
fallback onto contract law.  If the ARL OSL can't accomplish that, or worse, 
manages to invalidate itself in some manner, then that is a serious problem.

> Direction (2) - from a "contributor" to USG and the world, via "intentional 
> submission for inclusion in the Work" (clause 5).
>
> This direction (2) doesn't depend on copyright, nor on direction (1).
> In order to argue that
> it applies, at least in many cases, I'd suggest you need a good 
> contract/license, one that would stand when someone "intentionally
> submits for inclusion in the Work to the Licensor". Because with a 
> license/contract, ANY work "intentionally submitted" is licensed by its
> author under ARL OSL.
>
> Including patent grant, and things like the case you pointed out on 
> wikipedia shouldn't happen.

Agreed!

> >> But I'm not sure what you're worried about, sue for what? These
> >> (a)-(d) obligations have nothing to do with suing users, do they? ARL
> >> OSL has all the other clauses, which apply fine regardless of whether
> >> the underlying Work is copyrighted or not, like disclaimers of
> >> liability and clause 5.
> >
> > No, the problem is that removing those terms suggests that you can
> > strip out the ARL OSL from any part that is in the public domain.
> > Once that happens, the material no longer has the ARL OSL protecting
> > downstream users from predatory and unscrupulous individuals.  That's all.
>
> ARL OSL would apply for the projects under ARL OSL, including
> *contributors* to these projects.
>
> It wouldn't apply if I take the software and reproduce it on my own (without 
> contributing anything anywhere).

OK, so what you're saying is that some person A can download the software and 
redistribute it to some person B.  A is under contract to the USG by the ARL 
OSL, but since A didn't make a contribution, the contract terminates with that 
person, leaving B free to sue over whatever claims they wish to make, correct? 
I don't see how B can successfully sue A, because A can just point to the USG 
stating that the USG was responsible for the code.  If B then sues the USG, 
the USG will point at the ARL OSL which covers itself and all contributors. If 
a particular contributor took B's copyrighted or patented work without 
permission, then that contributor is at fault (the USG itself could be a 
contributor, so it may have to defend itself in court as well).  B may be able 
to sue over liability, but I'm not sure how that would play out (another 
question for ARL Legal).

> It wouldn't apply if I make a derivative work and try to fool my downstream 
> by, I don't know, revoking the license I claimed to give them.
> But that's also true of Apache License 2.0.

So the ARL OSL is compatible with Apache 2.0 in that way too! ;)

Thanks,
Cem Karan

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5559 bytes
Desc: not available
URL: <http://lists.opensource.org/pipermail/license-discuss_lists.opensource.org/attachments/20160817/07bd21c9/attachment.p7s>