[License-discuss] BSD 3-clause and copyright notices

[Zluty Sysel]

Hi there,

Thanks for the reply.

On Wed, Sep 30, 2015 at 11:12 AM, Philippe Ombredanne
<pombredanne at nexb.com> wrote:
> On Tue, Sep 29, 2015 at 6:13 PM, Zluty Sysel <zluty.sysel at gmail.com> wrote:
>> Hi all,
>>
>> After conquering many hurdles along the way, it looks like the company
>> I am a part of is willing to release a good part of the source code we
>> own as open source software. Before we do that though there are a
>> couple of outstanding issues that I was hoping someone on this mailing
>> list could clarify. We want to use a BSD 3-clause and immediately
>> publish the source code on a public code repository allowing
>> contributions from users. The questions that have arisen are the
>> following:
>>
>> 1) Clause 2 requires users that distribute the software in binary form
>> to reproduce the copyright notice. Since the holder of the copyright
>> notice is the very same company that makes the source code available
>> to them, would it be possible to selectively waiver this obligation to
>> a particular set of users without infringing on the Open Source
>> definition or the BSD license itself?
>> If the answer was negative, would including the existence of such a
>> waiver in the license itself preclude it from being considered an open
>> source software license?
>
> As the copyright holder you can do as you please.
> You do not need to attribute yourself for your own code, though that
> is of course nice anyway ;)
> No need for a waiver.
> A waiver would be rather an awkward and weird thing.

Well, since the code is distributed under a license that requires
disclosure there was some fear internally that disclosure and license
reproduction would be mandatory for all users of the source code.

>
>> 2) When accepting contributions to the source code repository from
>> external sources, I have seen that is sometimes customary to include
>> an additional copyright line to the license text included at the top
>> of the source file, crediting the person or company that contributed
>> the new code or file.
>
> Either that or an author file. Simpler is better.

That's good to know. An author file would make our lives easier and we
have no problem adding one, the problems come later with the inclusion
of the BSD notice for binary distributions as I'll explain below.

>
>> Would then the waiver mentioned in question 1) be in violation of the
>> additional copyright holder(s)' rights?
>
> May be not in violation of their rights but in contradiction with your
> eventual obligations.
>
> Think about it this way:
> What if you were such an external contributor: you worked hard to
> provide code enhancements to this project.
> And as a thank you note, you have ..... nothing. This would not be
> great, would it?

We absolutely do not want to hide the contributors from the source
code in the public repository. The fact of maintaining and AUTHORS
file with a list of all contributors is not only acceptable, but
regarded as a bonus guarantee for our users that other eyes have
perused the code and not have found any significant shortcomings.
The problem comes with acknowledging the usage of this codebase in
binary distributions. Some of the future users of this source code are
also our current customers, and some of these customers do not want to
reveal that they are using our particular libraries for their (binary
distributed) product. Given this, let me rephrase: Can we allow these
customers not to reproduce the BSD license text even if our AUTHORS
file contains names and email addresses of people outside of our
company? Because that's really all we're after here, allowing certain
customers not to have to mention that they are using our libraries.


>> 3) When reproducing the copyright notice in binary distributions, must
>> one parse all source code files to find out all of the contributors'
>> names and include them in full? Or is it enough to simply provide a
>> LICENSE file that only credits the original author (the company that
>> made the source code available originally) so that users of the source
>> code can simply reproduce that particular file in their binary
>> distributions?
>
> This is your call. Projects often create an AUTHORS file to list
> contributors to keep things simple.
> And/or list the major contributors in the a LICENSE or COPYING.
> Again giving credits to contributors is the _right and nice thing_ to do.
> (check out the scancode-toolkit if you want to create such a list of
> copyrights, disclaimer: I am an author of it)

As mentioned above, giving credit to all contributors _in the source
code_ is something we want to do. It's the binary distributions that
we're wary of.
Thanks for the tip, we'll take a look but if we already start off with
an AUTHORS file then we might not need to parse the source code at
all. Instead we'll ask contributors to add themselves to that file.

>
>> Thank you in advance,
>
> In summary, my 2 cents: The BSD license is simple, so keep things simple.
> You do not need to credit yourself in your own redistribution.
> Forget about adding waivers or other weird things to it: the weirder
> your license, the less likely anyone will want to contribute anything.
> I would not touch code with such a weird waiver (even with a very long
> pole).
> And If you are lucky enough to ever receive contributions from others,
> giving credit whether required or not is _always_ the right thing and
> the nice thing to do.

I agree, and in fact this is the preferred solution among the
engineers involved in this at the company. We want to keep it simple
and use an established open source license, we just want to make sure
that customers that choose not to disclose that they're using our
software can continue to do so.

> Embrace open source and be happy!

Fingers crossed!

Thanks!

Zluty