Open Stack

Richard Fontana <rfontana at redhat.com> writes:

> On Tue, May 14, 2013 at 01:53:19PM -0500, Mark Brown wrote:
>> 
>> However, I have had some very *practical* experience in this area; I was 
>> at
>> one time directly involved in a major lawsuit surrounding open source, 
>> copyright
>> and licensing (see my email address & your first guess will probably be 
>> correct).
>> In several cases, the provenance of individual files was being 
>> investigated,
>> and per-file copyright statements were an important part of that.
>
> In the case of OpenStack, like many other modern projects, the most
> accurate file provenance record would seem to be the git commit
> history. That does not necessarily tell you anything conclusive about
> copyright ownership, but it is a better record to go by than examining
> copyright notices in source files (which, as noted, could well be, or
> become, inaccurate).

It's worth noting that while git is very helpful in this regard, it's
only as accurate as the people using it choose to make it.

For instance, if you copy a file from another source into a repository,
git will indicate that you are the author of every line in that file.
Of course, the provenance of that file can be traced to the commit,
where ideally, you indicated where (and when!) the file came from, or if
you didn't, well, someone can ask you and perhaps you'll remember.

In some projects, this doesn't happen very often, if ever.  However, in
OpenStack, we have a habit of moving large chunks of code around to
different repositories.  We slice off bits of projects into new ones, or
combine different projects together.  When I'm involved, I try to
preserve as much git history as possible, but even so, it's inevitable
that someone says, "Oh, I forgot I needed this file too, I'll just copy
it over."

Moreover, there's an entire sub-project, Oslo, tasked with the goal of
making copying files between our different OpenStack projects as
efficient as possible.  So this happens a lot.

In the case of Oslo, at least, with some working knowledge of the
project, the original authorship is probably not too hard to find (e.g.,
this file in cinder was introduced by Charlie in a commit that said
"Sync from oslo", so I should check for the same file in oslo at the
same time as that commit and I see it was authored by Bob in a commit
that said "Import from nova" so I should look for a similar file in nova
at that time and I see it was originally authored by Alice.  Easy.)

And then, as you point out, you can ask the author who owns the
copyright.

-Jim