Gentoo Archives: gentoo-project

From: Ulrich Mueller <ulm@g.o>
To: gentoo-project@l.g.o
Subject: [gentoo-project] Re: [gentoo-dev-announce] Poll: Would you sign a Contributer License Agreement?
Date: Thu, 31 May 2018 10:23:40
Message-Id: 23311.52516.719358.967392@a1i15.kph.uni-mainz.de
In Reply to: [gentoo-project] Re: [gentoo-dev-announce] Poll: Would you sign a Contributer License Agreement? by Ulrich Mueller
1 >>>>> On Thu, 31 May 2018, Ulrich Mueller wrote:
2
3 > (hoping that he hasn't snipped too much of the context)
4
5 Presumably I have. So find Greg's two messages below, in full (which
6 had gentoo-project in CC).
7
8 Ulrich
9
10
11 >>>>> On Wed, 30 May 2018, Greg KH wrote:
12
13 > On Wed, May 30, 2018 at 04:36:09PM +0200, Ulrich Mueller wrote:
14 >> A while back I requested information on past copyright assignments [1].
15 >> Since then, we have located some 30 of the assignment forms, signed by
16 >> developers (most of them retired by now) in 2004.
17 >>
18 >> Here is the second part of the exercise. The current draft of the new
19 >> Gentoo copyright policy [2] arranges for two procedures:
20 >>
21 >> 1. Certifying agreement to a "Gentoo Developer's Certificate of Origin"
22 >> by including a "Signed-off-by" line with every commit. This would
23 >> be virtually identical to the procedure used for the Linux kernel,
24 >> and would be mandatory. A draft of the Gentoo DCO can be seen at [3].
25
26 > Please please please do not "fork" the DCO. It was specifically
27 > designed so that any project can use it, as-is, with no changes needed.
28 > Yes, some foolish projects have gone off and rewritten it, but that was
29 > crazy, and they now wish they did not, as it requires corporate lawyers
30 > to manually have to go review the "new" document to ensure that it
31 > really is doing what it thinks it is doing.
32
33 > Again, please just use the DCO. It's at it's own web site, and is good
34 > to be used that way:
35 > https://developercertificate.org/
36
37 > Also, note, that if you do decide to copy it, I personally am going to
38 > get upset as it is a blatent copyright violation. So there is that
39 > issue...
40 > Hint doing a s/open/free/ on the original text does not mean that you
41 > suddenly have created a brand new document with no requirement to abide
42 > by the original document's copyright. I see you claim that it was
43 > published in 2005 with a CC-BY-SA-2.5 License? Do you have any
44 > reference for that, I know I spent a lot of time working on this in the
45 > past and I do not remember that...
46
47 > Again, just use the DCO, please.
48
49 >> 2. In addition, according to the current policy draft, developers would
50 >> be encouraged to sign a "Gentoo Contributor License Agreement (CLA)".
51 >> Its current draft version is at [4]. However, this would be
52 >> completely voluntary and *not* be required. The exact workflow
53 >> hasn't been drafted yet, but PGP signing of the form would be one
54 >> possibility. (Also note that the form includes fields for real name
55 >> and postal address.)
56 >>
57 >> The goals of the second item is to "make compliance with this policy
58 >> easier (fewer copyright holders to list), and allow the Foundation to
59 >> enforce copyrights and re-license content if appropriate" [2].
60 >> Apparently, we will only be able to achieve these goals if a
61 >> significant fraction of contributors will sign the CLA.
62 >>
63 >> So, before I pursue more work on the CLA I would like to ask all
64 >> developers and contributors:
65 >>
66 >> - Would you sign a "Gentoo Contributor License Agreement", similar to
67 >> the current draft in [4]?
68
69 > No, I personally will not sign any CLAs, sorry.
70
71 > Sent publically as the DCO thing should be discussed in public.
72
73 > thanks,
74
75 > greg k-h
76
77
78 >>>>> On Thu, 31 May 2018, Greg KH wrote:
79
80 > On Wed, May 30, 2018 at 11:44:34PM +0200, Ulrich Mueller wrote:
81 >> >>>>> On Wed, 30 May 2018, Greg KH wrote:
82 >>
83 >> > Please please please do not "fork" the DCO. It was specifically
84 >> > designed so that any project can use it, as-is, with no changes
85 >> > needed.
86 >>
87 >> We simply cannot. We have files in the Gentoo repository that are not
88 >> under a free software license, and for these we need an extra clause.
89
90 > Your "extra clause" is pretty odd. You took out the c) clause of the
91 > original DCO for some unknown reason as well, which is going to cause
92 > you big problems.
93
94 > Was this vetted by a lawyer? Again, this is going to cause companies
95 > to have to spend lots of time and money to be able to get anyone to use
96 > this, do not change things lightly.
97
98 >> Otherwise we would have to specify in the policy that certain commits
99 >> are excepted from the requirement of a Signed-off-by line, and IMHO
100 >> that would be a much worse solution.
101 >>
102 >> Addition of the extra clause for licenses and similar files resulted
103 >> from a long discussion on 2018-01-25 in the #gentoo-council channel,
104 >> which included three council members and a trustee.
105
106 > No license lawyers?
107
108 > Are you _sure_ you need this change?
109
110 >> > Yes, some foolish projects have gone off and rewritten it, but that
111 >> > was crazy, and they now wish they did not, as it requires corporate
112 >> > lawyers to manually have to go review the "new" document to ensure
113 >> > that it really is doing what it thinks it is doing.
114 >>
115 >> > Again, please just use the DCO. It's at it's own web site, and is
116 >> > good to be used that way:
117 >> > https://developercertificate.org/
118 >>
119 >> > Also, note, that if you do decide to copy it, I personally am going
120 >> > to get upset as it is a blatent copyright violation. So there is
121 >> > that issue...
122 >>
123 >> How is it a copyright violation? We create a modified version of
124 >> a document that was released under a Creative Commons Attribution-
125 >> ShareAlike 2.5 License. Distribution of modified versions is allowed
126 >> under this license, and I believe that we include proper attribution.
127 >> Also section 4b of CC-BY-SA-2.5 explicitly allows distribution of a
128 >> modified work under CC-BY-SA-3.0.
129
130 > Fair enough, but please be sure to run the fact that you are changing
131 > something is obviously copyrighted by someone else with a declaration
132 > that it can not be changed, by relying on the wayback machine to make
133 > that change past a copyright lawyer. There is a reason that the DCO is
134 > not under such a license anymore, as this "respin" proves it :)
135
136 >> > Again, just use the DCO, please.
137 >>
138 >> See above, the simple reason is that we need an exception for license
139 >> files.
140 >>
141 >> Then again, Linux might profit from such a clause too. See for example
142 >> the following commit:
143 >> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/LICENSES/preferred/GPL-2.0?id=255247c2770ada6edace04173b35307869b47d99
144 >>
145 >> The commit message carries two Signed-off-by lines (and a Reviewed-by
146 >> by yourself). But let's look what the document says about its license:
147 >>
148 >> + Everyone is permitted to copy and distribute verbatim copies
149 >> + of this license document, but changing it is not allowed.
150 >>
151 >> Clearly, this isn't an open source license, because it doesn't allow
152 >> modifications. So I wonder how the committer could certify agreement
153 >> to the DCO 1.1 there?
154
155 > Section b) should cover this nicely. If your lawyers somehow feel it
156 > does not, I will be glad to consult with the LF lawyers about this and
157 > have them discuss the matter.
158
159 > Also note that I really doubt that the fact that you can include
160 > verbatim copies of a license in a repo is going to make anyone upset at
161 > all, unless you modify that license text. So you might all be worried
162 > about nothing "real" at all here. License files are not code, just like
163 > documentation is not code, and almost all open source licenses do not
164 > cover either of them well, if at all.
165
166 > As an armchair thought experiment of this, how would the overall license
167 > of a GNU project's tarball release such as bash, which is GPLv3, cover
168 > the license file of the GPLv3 text that is included in the tarball?
169 > Would the inclusion of a file in the tarball that is obviously not under
170 > a free software license cause that project's license to somehow not be
171 > "free software"?
172
173 > It's a fun rabit hole to go down, but one that I think you will have to
174 > do on your own :)
175
176 >> > No, I personally will not sign any CLAs, sorry.
177 >>
178 >> This is interesting, since you had previously signed the copyright
179 >> assignment form to Gentoo Technologies, Inc. (To be precise, you PGP
180 >> signed it and sent it to recruiters@g.o on 2004-03-08.)
181
182 > That was because I was forced to do so in order to become a Gentoo
183 > developer at the time, and my employer at the time also insisted on it,
184 > as they were the owners of all of the work that I did on Gentoo, not me.
185 > I had no say in the matter at all, just like almost all other people
186 > employed in the US due to the standard employment contract used. So to
187 > be clear, that was not _me_ giving up any copyrights, it was my
188 > employer.
189
190 > My position has changed on how best to handle copyrights in the 14 years
191 > since then, and I currently am employed by someone who allows me to keep
192 > my personal copyright (while also giving them a copy) so I guess I
193 > should figure out how to somehow retroactively not-sign it :)
194
195 > Any hints as to how to do that?
196
197 > Anyway, my strongest suggestion as to why not to change to use your
198 > custom license is the fact that you will now require all Gentoo
199 > developers who work for companies that allow their employers to
200 > contribute to Gentoo, to now have to have their lawyers read over this
201 > new license and come to an understanding of it before those people are
202 > allowed to contribute. That's a huge waste of time and money and will
203 > make those companies, and developers, grumpy.
204
205 > And if developers ignore the fact that they should have run this change
206 > by their employers, that could get them into big trouble later on.
207
208 > thanks,
209
210 > greg k-h