Re: logwatch: list of copyright holders
>> The most important one is that not having all of the copyright holders
>> represented means that we don't actually know what terms we are able
>> to distribute the final work. A component of a work which is
>> unlicenced makes the entire work undistributable.
I don't understand why this is solved by a list of copyright holders. I
always assumed it is solved by clear licensing terms. Please give me a
clear argument, to ensure that I can convince upstream.
> IANAL, but I don't think so, or better, I don't agree to
> one assumption.
>
> Simple patches are not copyrightable (so FSF doesn't
> require copyright transfer).
>
> IMHO the patches sent to a upstream author which
> doesn't patch the original copyright (adding a name or
> a copyright line) should be interpreted as the above case.
> IMHO the author implicit acknowledges that the patch is
> simple and doesn't include enough intellectual work.
> So I interpret the patch as outside copyright laws
This assumption might be safe under US copyright law (which I don't know
at all), the problem is just that under some jurisdictions, you can't
disclaim copyright at all. If a work is copyrightable is to decide by a
judge, based on not really specific guidelines in the law. (And you
won't know until there are disputes, otherwise no judge will decide)
On the other hand if some author, with or without copyright notice in
the source code, later turns up and says, I'm copyright holder and I
didn't give permission to redistribute, the judge will probably laugh at
him and tell him, that he gave implicit permission by sending a patch to
the project maintainer. But IANAL.
> So from CVS (as you mentioned) should give an idea
> if such contributions are outside the copyright laws.
>
> For new files the situation is clearly different.
>
> How did the upstream author find the patch and files
> of other authors?
> I assume that you and other send patches to CVS or to
> the upstream author for inclusion.
Yes, that should be the case. The address for patches is not a mailing
list, so it might be hard to verify all patches, even with upstream
support. (Some initial commits of scripts have an empty comment in CVS,
and no copyright notice, like [1])
> If the author included code from other project, the license of
> the imported code should be know.
Yes, I don't expect a problem here, because all scripts are in fact
plugins that are tied to the architecture of logwatch, so they were
specifically written for logwatch.
Willi
[1] http://www2.cvs.logwatch.org:81/index.cgi/logwatch/scripts/services/afpd
Reply to: