1 |
On Sat, Oct 24, 2015 at 8:33 AM, Sven Vermeulen <swift@g.o> wrote: |
2 |
> |
3 |
> This was brought forward when we started accepting user contributions |
4 |
> through GitHub. Yes, we've had the discussion that we don't depend on it. |
5 |
> But now the question was how do we need to interpret "depend on"? |
6 |
> |
7 |
|
8 |
The Council took up this topic last week. I'm not suggesting the |
9 |
Trustees are bound by this, but they certainly should be informed by |
10 |
it. What we agreed upon was: |
11 |
|
12 |
"The Gentoo council encourages contributions to Gentoo via manyfold |
13 |
ways. However, it also recognizes that the usage of Github, being a |
14 |
closed-source service, poses the danger of data lock-in and should not |
15 |
be preferred. The question has been posed whether the current usage of |
16 |
Github is in line with the Gentoo social contract- a question still |
17 |
open to interpretation. |
18 |
With this background the council asks for implementation of |
19 |
* the two-way mirroring of Github pull requests to bugzilla (including |
20 |
comments and patches) |
21 |
* the public archiving of Github repository e-mail notifications |
22 |
* and the mirroring of Github pull request git branches on Gentoo |
23 |
infrastructure |
24 |
or functionally equivalent alternatives. The council believes that |
25 |
this should suffice for all developers to dispell doubts about |
26 |
adherence to the Gentoo social contract." |
27 |
|
28 |
There is another side to this discussion which hasn't really been |
29 |
touched upon. Even if we wanted to move away from Github, what could |
30 |
we actually do to prevent its use? I think that attempting to do so |
31 |
would be fairly divisive. You can't tell somebody what tools they can |
32 |
use to prepare their patches any more than you can tell them what text |
33 |
editor to use to author their ebuilds. A file that spent some time in |
34 |
Github looks the same to Gentoo as a file that did not. I don't think |
35 |
anybody is going to support kicking devs who use it, or even those who |
36 |
encourage its use for contributions. |
37 |
|
38 |
Likewise, we don't actually have a policy that forces devs to close |
39 |
bugs at all, so a dev could choose to only work on pull requests |
40 |
submitted in Github and ignore bugs in Bugzilla. I think that would |
41 |
be ridiculous and counter-productive, but strictly speaking it would |
42 |
be allowed by policy. And what is the alternative, forcing devs to |
43 |
close Bugzilla bugs in a certain time? We don't require such things |
44 |
because devs are volunteers and a dev closing two bugs is more useful |
45 |
to us than a dev who quits and closes zero bugs because he's being |
46 |
yelled at for not fixing twenty. |
47 |
|
48 |
I think we'll get further by encouraging collaboration however it |
49 |
happens. When I fix a bug in an openrc script it isn't because I |
50 |
personally benefit (I no longer use openrc), but rather because it |
51 |
usually isn't hard for me to do and I know that lots of others will |
52 |
benefit. So, when you get a bug reported in an unconventional way, by |
53 |
all means we should work to get it into Bugzilla, but we should be |
54 |
fixing bugs because they're bugs, not because of how they're reported. |
55 |
|
56 |
> |
57 |
> In the extreme case, could developers and users who contributed time and |
58 |
> effort to the Gentoo project ask for compensation the moment that we would |
59 |
> be in breach of the Social Contract? |
60 |
> |
61 |
|
62 |
Under the present system, I'd think no, because all the real |
63 |
contributions are licensed GPL-2+ which doesn't actually contain any |
64 |
of our social contract terms. If we accumulated some big war chest of |
65 |
donations and then used them contrary to our announced purposes, that |
66 |
might be grounds for a lawsuit. However, we're not spending any |
67 |
Foundation money on services like Github, and it seems unlikely that |
68 |
we'd ever choose to do so, precisely because of the social contract. |
69 |
While I think we shouldn't be opposed to developers using Github we |
70 |
shouldn't be funding it. |
71 |
|
72 |
I actually have suggested that Gentoo move towards something like the |
73 |
FSFe FLA for copyright, and that does actually contain some clauses |
74 |
for taking back contributions if Gentoo were to stray. However, that |
75 |
is targeted more at re-licensing. For example, if you gave Gentoo an |
76 |
exclusive license to your contribution under the FLA and Gentoo chose |
77 |
to re-license it under a proprietary license, then the license would |
78 |
be terminated and copyright would revert to you. However, even that |
79 |
approach doesn't cover the social contract, and Gentoo would still |
80 |
have the same rights towards contributed code as it has under the GPL. |
81 |
|
82 |
IMO, trying to build stuff like this into actual software licenses is |
83 |
unwise. As we can see there is a lot of debate over just what "depend |
84 |
on" means and that isn't really a good foundation for a legal |
85 |
document. I think the debate is healthy, but taking this into courts |
86 |
and bankrupting the Foundation over the issue is not. |
87 |
|
88 |
> So the second question is, what are the ramifications towards the Gentoo |
89 |
> community, Gentoo project and even Gentoo Foundation when Gentoo would be in |
90 |
> breach of this part of the Social Contract? |
91 |
|
92 |
I think the Social Contract is more about what we stand for. Unless |
93 |
we were to take this to an extreme, I doubt any court would want to |
94 |
touch it from a legal perspective. |
95 |
|
96 |
I think the real impact is that the Social Contract is a big part of |
97 |
what brings us together. If we completely disregard it, I suspect |
98 |
we'd see a lot of people drifting away. After all, we're all donating |
99 |
our time. We want to donate that time towards something that means |
100 |
something. So, the Social Contract is critically important to Gentoo |
101 |
regardless of whether it has any legal basis. |
102 |
|
103 |
There is always going to be some edge case that raises a vigorous |
104 |
debate. I think the key is that we're having this argument over a |
105 |
gray area that is really on the periphery of what we do. I don't |
106 |
think that means we're compromising our core values - the fact that |
107 |
we're so divided actually suggests to me that we take such matters |
108 |
very seriously. |
109 |
|
110 |
-- |
111 |
Rich |