| 1 |
On Mon, Oct 24, 2016 at 7:31 PM, Matt Turner <mattst88@g.o> wrote: |
| 2 |
> |
| 3 |
> Yeah, it seems to be painful no matter what you do (CLA, copyright |
| 4 |
> assignment, listing copyright holders) just in different ways :) |
| 5 |
> |
| 6 |
|
| 7 |
Well, the advantage of assignment is that it does simplify copyright |
| 8 |
tracking, since you own the copyright on everything. |
| 9 |
|
| 10 |
The problem is that it potentially cuts out a lot of contributions. |
| 11 |
There is also the problem of nations that do not allow assignment |
| 12 |
(though that could also be fixed in theory by just ending the Gentoo |
| 13 |
German conspiracy). |
| 14 |
|
| 15 |
I personally tend to favor a mandatory DCO (we absolutely need to know |
| 16 |
the copyright status of our code), and a voluntary FLA (which I tend |
| 17 |
to prefer to outright assignment as I think it lines up well with our |
| 18 |
always-free social contract). |
| 19 |
|
| 20 |
The issue remains of what to do with the copyright notice. I |
| 21 |
suggested just having enough names on the line to account for 51% of |
| 22 |
the code, which as far as I can tell is completely legal. That |
| 23 |
doesn't preclude just listing all the names (which is ugly, but |
| 24 |
administratively simple). But, that does help cover us in cases where |
| 25 |
we have some ebuild where we can only account for 60% of it. It also |
| 26 |
allows us to borrow anything from any other project that already has |
| 27 |
its copyrights well-documented. |
| 28 |
|
| 29 |
FYI, one of the original sparks that drove some of this thinking was |
| 30 |
the eudev copyright fiasco (which like a lot of systemd-related stuff |
| 31 |
was blown out of proportion IMO with things being attributed to malice |
| 32 |
which were simply a lack of thinking things through). This was the |
| 33 |
first time Gentoo really forked and internalized a major external |
| 34 |
project, and there was a clash between our previous practices designed |
| 35 |
for dev-written code and the large import of external code. Under the |
| 36 |
draft copyright policy we'd have just maintained the previous |
| 37 |
copyright headers, perhaps just reformatting them to the top of the |
| 38 |
file per our convention if they weren't already there (which is |
| 39 |
completely legal). If at some point enough code in a file got |
| 40 |
rewritten to have majority-FLA authorship we'd have the option to |
| 41 |
change it to Gentoo and others, though we wouldn't have to. |
| 42 |
|
| 43 |
-- |
| 44 |
Rich |
Archives