Open Stack

On Fri, May 17, 2013 at 08:57:32AM -0400, Russell Bryant wrote:
> On 05/17/2013 08:16 AM, Sean Dague wrote:
> > I'm generally +1 on this, however it's going to be hard to get
> > organizations to agree to this while other orgs are listed in the source
> > files. From an IBM perspective, I've gotten agreement that we're cool
> > with this, as long as it's consistent. A patch work where some
> > copyrights are still listed, but others aren't allowed in, isn't going
> > to fly.
> > 
> > My suggestion, is that we declare a flag day (like June 15), and that
> > someone needs to make an objection by then, otherwise this is new
> > policy. We create a new hacking rule that projects can use to enforce it.
> > 
> > Then we have a couple of volunteers lined up to generate, and review
> > through mass removals from the code for each openstack/ project, plus
> > flipping on the hacking rule.
> 
> +1 on the need for consistency.
> 
> In order of my personal preference:
> 
> 1) completely accurate and up-to-date copyright headers
> 2) no copyright headers (just the license)
> 3) what we have now (incomplete, inaccurate, out of date)
> 
> I'm fine with 1 or 2 ... but big +1 on not 3.
> 
> If we're going to #2, I definitely think we just need to remove all of
> it.  If we can't do that, we should just aim for #1 by doing a better
> job of documenting and educating on the expectations here, and having
> reviewers do a better job of checking for it.

I don't think #1 will ever succeed unless you find some way todo
automated checks of it. You'd need to record the corporate affiliation
of each developer in OpenStack. Then each time a new change comes in,
you'd need to validate that either the developer's name or their
employeer's name is already listed in a copyright tag on each of
the source files changed.

Also the list of copyright tags will get very large quite quickly.
eg looking at git history, nova/virt/libvirt/driver.py has had
contributions from ~80 different authors, covering ~40 unique
domain names. So I would expect at least 40 lines of copyright
statements there if we were doing option #1. Given that it is
apparently not legally neccessary, what purpose would a huge
list of names serve ?

To me the important thing is that every file holds a header that
clearly identifies the license, and that the source file is part
of the openstack codebase (so its origins are clear if copied to
another project's codebase). Perhaps you could also add a generic
line "Copyright held by various OpenStack Community Developers"
but it wouldn't really do much from a legal POV. Ultimately the
GIT history is the only thing that can give you authoratative
information about a file's authorship & thus copyright.

So I'd vote in favour of option #2 with a generic header applied
to all files.

Daniel
-- 
|: http://berrange.com      -o-    http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org              -o-             http://virt-manager.org :|
|: http://autobuild.org       -o-         http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org       -o-       http://live.gnome.org/gtk-vnc :|