Gentoo Archives: gentoo-project

From: "Michał Górny" <mgorny@g.o>
To: gentoo-dev@l.g.o
Cc: gentoo-project <gentoo-project@l.g.o>
Subject: [gentoo-project] Re: [gentoo-dev] [RFC] Splitting developer-oriented and expert user mailing lists
Date: Sun, 03 Dec 2017 22:04:09
Message-Id: 1512338638.22374.46.camel@gentoo.org
1 W dniu sob, 02.12.2017 o godzinie 19∶33 -0600, użytkownik R0b0t1
2 napisał:
3 > Hello,
4 >
5 > In every mailing list conversation, there are at least three people:
6 > the two conversing, and the future reader. I point this out as I think
7 > it important that everyone realize that not all posts are written for
8 > those immediately participating in the conversation.
9 >
10 > Some time ago I was offered some equipment due to my history of
11 > open-source contributions to a variety of projects. I asked the donor
12 > to forward it (or money) to the Gentoo foundation, but they declined,
13 > citing a general distaste for the management of software projects in
14 > general and specific issues they believed existed within Gentoo.
15
16 I'm not sure if this is relevant to the topic at hand. There are many
17 issues within Gentoo. I'm trying to address one of them.
18
19 > On Sat, Dec 2, 2017 at 5:18 PM, Michał Górny <mgorny@g.o> wrote:
20 > > Hello, everyone.
21 > >
22 > > This is something that's been talked about privately a lot lately but it
23 > > seems that nobody went forward to put things into motion. SO here's
24 > > a proposal that aims to improve the condition of our mailing lists
25 > > and solve some of the problems they are facing today.
26 > >
27 >
28 > If you have in fact discussed this off list with people who agree, I
29 > think it is important that you invite them to comment. Not only will
30 > it show support for what you have detailed, it will allow them to
31 > explain the problems they have in greater detail, so that perhaps a
32 > solution that does not involve restricting list access could be found.
33
34 This sentence merely focuses on 'don't shoot the messenger' part which
35 will happen anyway. Those people won't come here to '+1' the proposal
36 because this mailing list is not supposed to be about mail popularity
37 contests.
38
39 Also because they don't want to be targeted by people misbehaving here.
40 In fact, a number of them already pinged me today privately showing
41 support, and some of them told me exactly that -- that they don't want
42 to become a target of aggression. A few participants of this mailing
43 list have shown harassment towards people that stood up to them --
44 including constant insults on various public and private channels.
45
46 >
47 > It may be that I am misunderstanding your language, but what you have
48 > presented does not leave many things open for discussion. It seems
49 > like what you have presented is to be either accepted or rejected as
50 > is. Seeing as my opinion does not matter, it further seems like it
51 > will simply be accepted as is.
52
53 I simply don't believe that after so many iterations there's any more
54 option that hasn't been tried or rejected already.
55
56 > >
57 > > Problems
58 > > ========
59 > >
60 > > Currently the developer-oriented mailing lists gentoo-dev and gentoo-
61 > > project are open to posting by everyone. While this has been generally
62 > > beneficial, we seem to be having major problems with some
63 > > of the posters for more than a year. Off hand, I can think of three:
64 > >
65 > > 1. Repeating attacks against Gentoo and/or Gentoo developers (including
66 > > pure personal attacks). While it is understandable that some people may
67 > > be frustrated and need to vent off, repeating attacks from the same
68 > > person are seriously demotivating to everyone.
69 > >
70 >
71 > No one has any right to not be offended. If Gentoo developers are
72 > receiving criticism for their behavior, then perhaps it would be best
73 > that they critically analyze their actions and the effect that they
74 > have on other people.
75 >
76 > As far as I am aware most developers never get harassed and go quietly
77 > on about their business. I have even asked some questions similar to
78 > the questions I have asked on this list that people have felt were
79 > adversarial. However, these developers didn't seem to mind my
80 > questions and spent 5 minutes or so of their time on a response.
81 >
82 > > 2. Frequent off-topics, often irrelevant to the thread at hand.
83 > > I understand that some of those topics are really interesting but it is
84 > > really time-consuming to filter through all the off-topic mails
85 > > in search of data relevant to the topic at hand. What's worst, sometimes
86 > > you don't even get a single on-topic reply.
87 > >
88 >
89 > Does the list have a digest subscription option? I find that extremely
90 > helpful for one list I am subscribed to (Perl6 development) which is
91 > very high volume. On the other hand, lots of offtopic chatter would
92 > still be hard to sort through, but I think it needs to be considered
93 > whether the chatter the list currently receives is truly off topic.
94 > What if it is simply concerns or subjects that the OP did not want to
95 > consider? Does that make it off topic? Is the problem more involved
96 > than previously thought?
97 >
98 > > 3. Support requests. Some of our 'expert users' have been abusing
99 > > the mailing lists to request support (because it's easier to ask
100 > > everyone than go through proper channels) and/or complain about bug
101 > > resolutions. This is a minor issue but still it is one.
102 > >
103 >
104 > In the case of actual support requests, it might be worth taking some
105 > kind of action against the user, but the general level of competence
106 > of Gentoo users makes me wary that this may be a mischaracterization
107 > of the intent of the email. If something like a "support request"
108 > percolates to gentoo-dev, it may be of a similar vein as a complaint
109 > about a bug resolution. Complaining about bug resolutions seems valid,
110 > especially if questions on the tracker have been ignored.
111 >
112 > Some developers in particular seem to not appreciate being held
113 > accountable for their actions. In most notable cases, all anyone ever
114 > does is ask for an explanation as to why something occurred - and in
115 > most notable cases, that question is ignored, with no recourse left to
116 > the user or contributor.
117 >
118 > Personally, I tried to ask why eix's "optimizations" flag was removed,
119 > when other packages *do the exact same thing.* Still no response. How
120 > am I supposed to interpret this?
121
122 I'm sorry but the purpose of this thread is not to convince you that
123 the problems exist. If you haven't experienced them already, then it
124 would be polite of you to either accept them as a fact, or do some
125 research yourself.
126
127 I understand that you might want to know things. However, it is
128 generally impolite if someone 'comes late to the party' and starts
129 shouting questions that the existing participants know answers to
130 already. This is distorting to the conversation at hand.
131
132 In such a situation, as I said it is usually polite to try to find
133 the answers yourself or politely and privately query one
134 of the participants who you are acquainted to or is otherwise able
135 and willing to help you.
136
137 >
138 > >
139 > > All of those issues are slowly rendering the mailing lists impossible to
140 > > use. People waste a lot of time trying to gather feedback, and get
141 > > demotivated in the process. A steadily growing number of developers
142 > > either stop reading the mailing lists altogether, or reduce their
143 > > activity.
144 > >
145 > > For example, eclass reviews usually don't get more than one reply,
146 > > and even that is not always on-topic. And after all, getting this kind
147 > > of feedback is one of the purposes of the -dev mailing list!
148 > >
149 >
150 > It may be that this is separate from the content of the mailing list.
151 > Do some of the developers simply not like the format of a mailing
152 > list? A lot of projects are now using Slack and Discourse in addition
153 > to IRC. I personally do not like either of those services, but some
154 > people think they allow reduce response times, aid in comprehension,
155 > allowing greater involvement of developers.
156 >
157 > As it is, it seems to me like a lot of development happens on IRC and off list.
158
159 Yes. Sometimes only because IRC is much faster. Sometimes because using
160 mailing lists becomes impossible due to problems listed above.
161
162 > >
163 > > Proposal
164 > > ========
165 > >
166 > > Give the failure of other solutions tried for this, I'd like to
167 > > establish the following changes to the mailing lists:
168 > >
169 > > 1. Posting to gentoo-dev@ and gentoo-project@ mailing lists will be
170 > > initially restricted to active Gentoo developers.
171 > >
172 > > 1a. Subscription (reading) and archives will still be open.
173 > >
174 > > 1b. Active Gentoo contributors will be able to obtain posting access
175 > > upon being vouched for by an active Gentoo developer.
176 > >
177 > > 2. A new mailing list 'gentoo-expert' will be formed to provide
178 > > a discussion medium for expert Gentoo users and developers.
179 > >
180 > > 2a. gentoo-expert will have open posting access like gentoo-dev has now.
181 > >
182 > >
183 > > Rationale
184 > > =========
185 > >
186 > > I expect that some of you will find this a drastic measure. However, I
187 > > would like to point out that I believe we've already exhausted all other
188 > > options to no avail.
189 > >
190 >
191 > There is an option that has not been discussed, and that is
192 > questioning why the gentoo-dev list receives offtopic replies,
193 > personal attacks, and trolling.
194
195 People's private issues are not topic of this mailing list. It is
196 generally impolite and unprofessional to discuss them publicly. Please
197 don't do that.
198
199 > > The problems of more abusive behavior from some of the mailing list
200 > > members have been reported to ComRel numerous times. After the failure
201 > > of initial enforcement, I'm not aware of ComRel doing anything to solve
202 > > the problem. The main arguments I've heard from ComRel members were:
203 > >
204 > > A. Bans can be trivially evaded, and history proves that those evasions
205 > > create more noise than leaving the issue as is.
206 > >
207 > > B. People should be allowed to express their opinion [even if it's pure
208 > > hate speech that carries no value to anyone].
209 > >
210 > > C. The replies of Gentoo developers were worse [no surprise that people
211 > > lose their patience after being attacked for a few months].
212 > >
213 >
214 > People only ever do things that make sense. Again, I invite the people
215 > who are being attacked to consider why someone cares enough to bother
216 > to do that. Bored teenagers go to #archlinux to have pissing contests,
217 > not #gentoo.
218 >
219 > >
220 > > The alternative suggested by ComRel pretty much boiled down to 'ignore
221 > > the trolls'. While we can see this is actually starting to happen right
222 > > now (even the most determined developers stopped replying), this doesn't
223 > > really solve the problem because:
224 > >
225 >
226 > To me this sounds like ComRel realized it is too easy to turn good
227 > intentions into fascism.
228 >
229 > > I. Some people are really determined and continue sending mails even if
230 > > nobody replies to them. In fact, they are perfectly capable of replying
231 > > to themselves.
232 > >
233 > > II. This practically assumes that every new mailing list subscriber will
234 > > be able to recognize the problem. Otherwise, new people will repeatedly
235 > > be lured into discussing with them.
236 > >
237 > > III. In the end, it puts Gentoo in a bad position. Firstly, because it
238 > > silently consents to misbehavior on the mailing lists. Secondly, because
239 > > the lack of any statement in reply to accusations could be seen
240 > > as a sign of shameful silent admittance.
241 > >
242 >
243 > It is also entirely possible that a new user will see the troll, agree
244 > with the troll, and not want to contribute to Gentoo because they
245 > think the troll is right.
246 >
247 > >
248 > > Yet another alternative that was proposed was to establish moderation of
249 > > the mailing lists. However, Infrastructure has replied already that we
250 > > can't deploy effective moderation with the current mailing list software
251 > > and I'm not aware of anyone willing to undergo all the necessary work to
252 > > change that.
253 > >
254 > > Even if we were able to overcome that and be able to find a good
255 > > moderation team that can effectively and fairly moderate e-mails without
256 > > causing huge delays, moderation has a number of own problems:
257 > >
258 > > α) the delays will make discussions more cumbersome, and render posting
259 > > confusing to users,
260 > >
261 > > β) they will implicitly cause some overlap of replies (e.g. when N
262 > > different people answer the same question because they don't see earlier
263 > > replies until they're past moderation),
264 > >
265 > > γ) the problem will be solved only partially -- what if a reply contains
266 > > both valuable info and personal attack?
267 > >
268 >
269 > I agree with this logic, but please be careful - it states a problem,
270 > presupposes a single solution, and then concludes that there is only
271 > one course of action based on the critique applied to that one
272 > solution. This is partly why I see the proposal as something which
273 > does not seem to be accommodating to alternate viewpoints. It makes
274 > addressing this section with an alternate viewpoint difficult, and if
275 > I ignore it then it looks like I ignored part of your argument.
276 >
277 > >
278 > > Seeing that no other effort so far has succeeded in solving the problem,
279 > > splitting the mailing lists seems the best solution so far. Most
280 > > notably:
281 > >
282 > > а. Developer mailing lists are restored to their original purpose.
283 > >
284 > > б. It is 'fair'. Unlike with disciplinary actions, there is no judgment
285 > > problem, just a clear split between 'developers' and 'non-developers'.
286 > >
287 > > в. 'Expert users' are still provided with a mailing list where they can
288 > > discuss Gentoo without being pushed down into 'user support' channels.
289 > >
290 > > г. Active contributors (in particular recruits) can still obtain posting
291 > > access to the mailing lists, much like they do obtain it to #gentoo-dev
292 > > right now. However, if they start misbehaving we can just remove that
293 > > without the risk of evasion.
294 > >
295 >
296 > I feel this is still a fairly large barrier to involvement. Getting
297 > people to the point they want to contribute or have the knowledge to
298 > contribute is the hard part, and what this will make harder to do.
299
300 A mailing list is not strictly essential to contributing to Gentoo.
301 I can't think of it being to much use of any recently recruited
302 developers.
303
304 --
305 Best regards,
306 Michał Górny