[PATCH weston] Update COPYING

[Bryce Harrington]

On Fri, Feb 17, 2017 at 05:19:55PM +0000, Daniel Stone wrote:
> Hi,
> 
> On 17 February 2017 at 17:12, Bryce Harrington <bryce at osg.samsung.com> wrote:
> > On Fri, Feb 17, 2017 at 07:50:45AM -0600, Derek Foreman wrote:
> >> How about we just remove all the copyrights from COPYING and let
> >> anyone that needs to collate this list do it themselves...
> >>
> >> That way we don't need to remember to update it in 2 weeks when
> >> someone adds a copyright to a file and doesn't realize we duplicate
> >> the info in COPYING? :)
> >
> > +1
> >
> > It seems atypical amongst open source projects to have an exhaustive
> > (and duplicative as pointed out) listing of copyright statements.  Has
> > there been an issue raised from outside the project that this listing
> > would solve?
> 
> The X projects all have it, at least.

I only spot checked, and you would know better than me, but the ones I
looked at appeared to just list key/major copyright holders (which does
seem sensible).  "Vague indications of copyrights", as you mention below.

> Everyone who distributes it
> (distributions, companies building products, etc) need to have
> something that at least minimally conforms to the Mesa licensing
> document: a full statement of the license, and at least a vague
> indication of the copyrights. Depending on the legal department
> involved, they may end up compiling this exact list for their own use.

I've done a couple such conformance checks in the past, and indeed I had
to compile such a list, so you're certainly right.  But as pq showed,
it's a straightforward set of shell commands to do it.  And actually, if
I were doing a conformance check, I wouldn't trust that the COPYING file
was being kept up to date so would do that scan regardless.  Indeed, if
there were any descrepancies that showed up I would feel compelled to
investigate each of them.  IOW rather than saving time the COPYING file
might actually create an bit of extra work for compliance checker.
Frankly, the script itself might be more valuable in this regard, so
maybe that's what should be included in the tree?

If the ultimate goal is to help make compliance checking easier, I would
suggest focusing on fixing any irregularities in the files themselves -
e.g. continuing to ensuring dates and copyright formats are correct,
that boilerplate licensing text is consistent across files, etc., as
folks have been doing, so that running a scan is clean and reliable.

OTOH, if the goal is about giving recognition to contributors, an
AUTHORS or CONTRIBUTORS file seems to be more conventionally used
approaches.

> I'd be fine to reduce it to the minimal license text, but that doesn't
> free us up from needing to check incoming source to make sure it
> conforms to the same license. We should really also merge data/COPYING
> into the core COPYING.

Obviously checking licenses on incoming code is always extremely
important. :-)

I'm not sure what you're suggesting by reducing it to the minimal
license, the file only includes one license statement so appears to be
minimal already; I'm not suggesting copyrights *shouldn't* be present,
or that any of the existing ones should be removed.  AIUI it's required
to have at least one copyright statement, and seems pretty standard to
list the major copyright holders (esp. any companies/individuals with a
legal interest.)  The main purpose of COPYING, though, is the licensing,
to document how the codebase can be shared and reused.

You're probably right that merging data/COPYING and COPYING makes
sense, but I've seen enough other projects that had subdir-specific
licensing gunk that I'm not really worried about it.  I'd be fine
either way.

Bryce