1 |
Hi, everyone. |
2 |
|
3 |
I'd like to request Infra to establish a new mailing list that would |
4 |
fill in the gap between our public mailing lists and the gentoo-core |
5 |
mailing list. |
6 |
|
7 |
Name: gentoo-dev-internal |
8 |
|
9 |
Topic: technical discussions between active Gentoo contributors |
10 |
|
11 |
Restrictions: |
12 |
|
13 |
- public (open subscription), initially we may optionally copy all |
14 |
subscribers from gentoo-dev so that they do not miss discussion, |
15 |
|
16 |
- archived, |
17 |
|
18 |
- but posting restricted to opt-in member group. |
19 |
|
20 |
Initially, the posting group would include active Gentoo developers |
21 |
only. Afterwards, we will deploy a small moderator team whose purpose |
22 |
would be controlling access to the list -- including both adding new |
23 |
members on request and removing existing members (including developers) |
24 |
if they misbehave. |
25 |
|
26 |
I don't think we need to precisely define the rules for admitting new |
27 |
members. I think the exact procedure would be at moderators' discretion |
28 |
and would depend on the current 'health' of candidates -- i.e. if things |
29 |
go calm they may just admit on request, and if people start abusing this |
30 |
they will force explicit moderation before whitelisting. |
31 |
|
32 |
|
33 |
Rationale |
34 |
========= |
35 |
|
36 |
The purpose of gentoo-dev is to allow technical discussion between |
37 |
contributors to Gentoo, especially including making it possible for |
38 |
developers to send RFCs and discuss their ideas. |
39 |
|
40 |
Sadly, it is not uncommon for threads on that mailing list to turn into |
41 |
trollfests, get deranged or hijacked into completely different topics. |
42 |
Things are so bad that the mailing list stops serving its purpose. It |
43 |
involves a number of consequences: |
44 |
|
45 |
a. The developers lose time on the mailing lists instead of using it for |
46 |
constructive purposes. Even skimming through those mails in search of |
47 |
something remotely relevant is time-consuming. |
48 |
|
49 |
b. The developers and contributors become discouraged and unsubscribe |
50 |
from the mailing list. As a result, audience for reviews and RFCs |
51 |
becomes smaller and even less focused on the topic. |
52 |
|
53 |
c. The developers become discouraged and stop sending their ideas. |
54 |
Either they do less, use another media or work in complete isolation |
55 |
from other community members. |
56 |
|
57 |
d. Eventually, the developers become tired of the persisting issues |
58 |
and they retire (yes, it's a fact). |
59 |
|
60 |
Other ideas on solving those issues were pretty much rejected already: |
61 |
|
62 |
1. Bans on persisting violators were rejected as they are easily worked |
63 |
around via subscribing from another e-mail address, and causing more |
64 |
noise than the original issue. |
65 |
|
66 |
2. Full-scale moderation of mail on gentoo-dev was rejected because of |
67 |
technical limitations and the resulting high level of effort in handling |
68 |
the moderation, plus the social effect. |
69 |
|
70 |
3. Making gentoo-dev@ opt-in (like the suggested new list) would make it |
71 |
much harder for users to contribute ideas, and would inevitably |
72 |
discourage some of the users from writing. |
73 |
|
74 |
All that considered, establishing a second mailing list with different |
75 |
characteristic seems like a reasonable solution. In particular: |
76 |
|
77 |
A. It gives a wider choice of tools for developers (and privileged |
78 |
contributors) -- they can choose either the open or restricted mailing |
79 |
list depending on the type of requested feedback. |
80 |
|
81 |
B. The gentoo-dev mailing list is still open for power users |
82 |
and contributors to submit their own ideas, and with no moderation |
83 |
the discussion can proceed naturally. |
84 |
|
85 |
C. The cost of moderation should be relatively low, and the methods can |
86 |
be dynamically adjusted to fit the needs. In particular, good behavior |
87 |
on gentoo-dev can be used to grant access to gentoo-dev-internal without |
88 |
further requirements. |
89 |
|
90 |
D. The restricted mailing list should be resilient to ban evasion since |
91 |
the access is opt-in, and the moderators team can enforce direct |
92 |
moderation of new members if there is a considerable risk. |
93 |
|
94 |
Your comments? |
95 |
|
96 |
-- |
97 |
Best regards, |
98 |
Michał Górny |