On Mon, Feb 16, 2015 at 7:44 AM, Joshua Kinard <ku...@gentoo.org> wrote:
>
> As far as removing the ebuild goes, that was probably the correct course of
> action, because we Yanks love to make our legal code as bizzaringly complex as
> we think we can.  Though, the mistaken code is still in CVS in the Attic --
> does that itself present any problems that need to be addressed?
>

So, there are a bunch of issues here, but let's just address whether
the copyright line was a problem and set aside all the
personal/organizational/procedural stuff:

1.  We have a clear policy that the copyright line must be exactly
foo, and this one wasn't.
2.  MAYBE violating that policy could or couldn't cause an issue, but
the legalities of that become really messy really fast, so it is
better to just follow the policy until it is changed.
3.  I don't really see a problem with having the file in the Attic
unless somebody asks us to take it down.  The file is legally
redistributable, after all.

A few of the issues I see with having the file in the tree unmodified:
1.  It is GPL-2, not GPL-2+, which could create issues with
relicensing if we wanted to.  If copyright were assigned to the
Foundation there would not be an issue with that.  Yes, I realize that
the current policy is at best ambiguous on that front.  However, we're
not doing ourselves any favors by switching from an ambiguous but
potentially advantageous approach to an unambiguous but
disadvantageous approach.
2.  It opens the door to lots of other situations like this in the
absence of any sane policy for dealing with them.

The current policy requires committers to ensure that it is legal to
put in the copyright line as the policy dictates, and to keep stuff
out of the tree if not.  It isn't a great policy, but it is at least
workable and obviously being the status quo it is what it is.

I do think that moving to a cleaner policy makes a lot of sense.  The
problem is that doing this sort of thing right potentially involves a
lot of work as well.  Maybe another approach is to just ditch per-file
copyrights entirely (which a random perusal suggests is how Linux does
things), but that would STILL require stripping the copyright out of
these files with all the issues that entails, and limit our ability to
borrow license-compatible code.

-- 
Rich

Reply via email to