Copyright for reworked Phobos code in Mir

Steven Schveighoffer schveiguy at gmail.com
Wed Dec 26 16:17:08 UTC 2018


On 12/26/18 10:40 AM, 9il wrote:
> Hi folks,
> 
> I am slightly confused by copyright mess in some of Mir modules. As you 
> may know, some of them contain reworked Phobos functions. Plus I am not 
> sure that I understand the meaning of Copyright in the context that both 
> Phobos and Mir are Boost licensed.

Disclaimer: I am not a lawyer.

> 
> For example, currently, I am creating mir.numeric that will contain 
> findRoot, findLocalMin rework of Phobos and other stuff. And 
> findLocalMin in Phobos is my work.
> 
> std.numeric contains:
> Copyright: Copyright Andrei Alexandrescu 2008 - 2009.

Copyrights headers in Phobos are a bit unmaintained. If someone writes 
any code, technically, they could assert copyright on their 
contribution. So really, the list should just be everyone who ever 
committed *new code* to the module (including yourself). Just moving 
stuff around, or fixing a simple bug probably isn't going to be 
copyrightable I think (reminder: I am not a lawyer).

But copyright notices aren't the final say in whether something is 
copyrighted by someone. It's really the evidence of who committed what 
that makes the difference.

> 
> What copyright should contain mir.numeric?

What I would do is copy that copyright notice into your code. Then at 
least you are conforming to the license. Under the terms of the boost 
license, there isn't much to assert claim to, and I would expect anyone 
who contributed to a module in a significant way who wants to get 
recognized there would easily get added to the module.

> 
> Another example is that sometimes I write a new implementation but use 
> Phobos unittests.

Code is copyrightable, even if it's unittests. So technically that 
portion should be boost licensed if you copied from Phobos (if it's not 
already). Under the terms of the boost license, you really only need to 
specify the license/copyright for the source code, there is no 
limitations on binary release.

In that case, I would transfer whatever copyright notice is on the file 
the contains the unittests.

> 
> There was an inverse precedent - Mersenne Twister. Mir version was 
> backported to Phobos.

The author (I'm assuming you) should get credit under the copyright 
header, if you are not, and it's important to you, just submit a PR.

I've contributed loads of different things all over the place in Phobos 
and druntime, I almost never add my name to the copyright line, because 
I'm not too concerned about it.

-Steve


More information about the Digitalmars-d-learn mailing list