On 31/01/13 10:20, David Woolley wrote:
> The purpose of such clauses is not to track the provenance, but to
> maintain the purity of the  official version, so that forks cannot be
> passed off as approved versions.

This sort of protection is the domain of trademark law, not copyright
law. Attempting to do it via copyright (while keeping the license open
source) leads to clunky requirements like this one.

> One of the key objectives of open source is code re-use.  This means
> that a file must be usable outside the context of the original
> application, and therefore with a different, or no, source management
> system.

And therefore, increasingly, the "official version" is determined by
community consensus and developer resources, not by the fact that it's
maintained by the guy who originally wrote it. You then end up with the
position where all the checkins on the version which everyone considers
canonical have to have such comments.

> The meta data also often doesn't contain legal identity of the
> modifieres, and doesn't distinguish between de mimis changes, and ones
> where the modify owns copyright.

There is absolutely no way you will ever maintain a state where everyone
who owns a copyright interest in a file has added a note to the top of
it, and everyone who does not own such an interest has not. As someone
who did a lot of work with the MPL 1.1, which sort of attempted this
kind of copyright tracking, I know.

> There may be compromises, but the less information you include in a
> file, the more difficult it is to re-license it, once divorced from the
> original source management system.

Unless you collect copyright assignments, you should assume that
re-licensing _anything_ will be difficult in the future, in-file
comments or not. Even if you have such comments, you cannot assume they
are accurate, and you have to do all the due diligence anyway. (Again, I
speak as the guy who did what is probably the largest ever relicensing
of an open source codebase with heterogenous copyright.)

License-discuss mailing list

Reply via email to