[D-runtime] druntime license and copypasta bugs

Sean Kelly sean at invisibleduck.org
Mon Sep 5 10:42:58 PDT 2011


On Sep 4, 2011, at 2:38 PM, David Nadlinger wrote:

> On Sep 3, 2011, at 11:31 AM, Jonathan M Davis<jmdavisProg at gmx.com>  wrote:
> > That's easy enough to fix, but who's the copyright owner if not the authors?
> > Digital Mars?
> 
> On 9/3/11 10:47 PM, Sean Kelly wrote:
>> For runtime and Phobos I think ts always the original author.  Not so with GCC code though, for example.
> 
> 
> First, I am not a lawyer, and indeed it would be great if we could get advise from an actual copyright expert on the subject. But still, a few ideas from my experience with other open source projects, like KDE:
> 
> As far as I know, there is no such thing as a single copyright holder or »original author« per source file, but instead all contributors retain the rights for the code they submit. Thus, the actual copyright for a source file in question is typically scattered among several authors. Because of this, e.g. files from the KDE project includes a copyright line for all authors, like:
> 
> Copyright <year>-<year> <name of first author> <first e-mail>
> Copyright <year> <name of second author> <second e-mail>
> 
> If I understand the situation correctly, requiring a single copyright line resp. making a difference between »Authors:« and »Copyright:« makes no sense for typical open source projects, but it could well be I am wrong here.

Interesting.  I had always assumed that the copyright holder on a file was the original author of the file, and everything else was just modifications to the original.  That's simply a completely uninformed assumption though.

> While we are at it, easily being able to reconstruct the ownership status and contact all the authors is one of the key reasons behind Git recording name and email from both author and committer. This obviously doesn't obviate any legal requirements for copyright notices or similar, but can be helpful in tracing down potential claims.

That's really very nice.  Tango is in license hell right now, for example, because it has so many contributors and few of them can actually be contacted.  There's no way for Tango to use a new license at this point.

> By the way, the copyright being split among many authors is the reason why it near impossible for larger projects (e.g. the Linux kernel) to change the license or to pursue license violations, because it would require obtaining approval from all the people who ever contributed a patch. For this reason, some projects require all contributors to assign their copyright to some central legal entity, usually a nonprofit.

Yeah… see above.  Since we use the Boost license (which is very open) I don't see this being an issue for us, as I don't foresee any need to change the license.  But it would certainly be good insurance.


More information about the D-runtime mailing list