1 |
05.02.2015 22:16, Michał Górny пишет: |
2 |
> Dnia 2015-02-05, o godz. 22:05:55 |
3 |
> Sergey Popov <pinkbyte@g.o> napisał(a): |
4 |
> |
5 |
>> 05.02.2015 21:22, Michał Górny пишет: |
6 |
>>> Dnia 2015-02-05, o godz. 21:08:43 |
7 |
>>> Mikle Kolyada <zlogene@g.o> napisał(a): |
8 |
>>> |
9 |
>>>> |
10 |
>>>> 05.02.2015 09:56, Michał Górny пишет: |
11 |
>>>>> Hello, everyone. |
12 |
>>>>> |
13 |
>>>>> It's finally time to discuss some of the recruitment issues. It's not |
14 |
>>>>> a new complaint that the process is time-consuming and discouraging to |
15 |
>>>>> our contributors. We have a pretty low number of new recruits [well, |
16 |
>>>>> we could definitely have a higher number!] and too often they resign |
17 |
>>>>> in the process. |
18 |
>>>>> |
19 |
>>>>> As I see it, the main issue are ebuild quizzes. They are very time- |
20 |
>>>>> consuming and discouraging. It's like filling a quiz with relatively |
21 |
>>>>> simple questions where answers need to fit a key, and you have to tell |
22 |
>>>>> the recruit to fill in the missing bits a few times just to help him |
23 |
>>>>> get further. |
24 |
>>>>> |
25 |
>>>>> I myself attempted ebuild quiz twice, because the first time I simply |
26 |
>>>>> ended up not having the time for it. My late recruit was making slow |
27 |
>>>>> progress, and recently vanished -- hopefully only because he doesn't |
28 |
>>>>> have will for that anymore. As I see it, the disadvantages outweigh |
29 |
>>>>> the benefits here. |
30 |
>>>>> |
31 |
>>>>> I have discussed this with kensington and a few Council members |
32 |
>>>>> (unofficially), and we came up with following ideas: |
33 |
>>>>> |
34 |
>>>>> 1. remove or reduce the ebuild quiz to a reasonable number of |
35 |
>>>>> questions. In other words, make it bearable. Focus on the stuff that |
36 |
>>>>> can't be checked otherwise. |
37 |
>>>>> |
38 |
>>>>> 2. Add an extra contribution period in which the candidate commits to |
39 |
>>>>> the tree through Pull Requests. Developers watch the requests, review |
40 |
>>>>> them and decide when the recruit is ready. We may extend this with |
41 |
>>>>> requirements like '3 different developers must review late activities |
42 |
>>>>> and evaluate them'. |
43 |
>>>>> |
44 |
>>>>> 3. Possibly extend the recruit-recruiter interaction. Rather than |
45 |
>>>>> treating the interrogation as some kind of final confirmation, make it |
46 |
>>>>> a small extra part of the learning process. In other words, reduce |
47 |
>>>>> the other parts, fill in the blanks here. |
48 |
>>>>> |
49 |
>>>>> What do you think? |
50 |
>>>>> |
51 |
>>>> I try to be short. |
52 |
>>>> First of all, i'm against to make quizzes simpler or shorter. I have few |
53 |
>>>> real examples, when people do a LOT of mistakes even after they passed |
54 |
>>>> the quizzes. |
55 |
>>> |
56 |
>>> And how people doing mistakes *after* passing the quizzes proves that |
57 |
>>> quizzes are good? As I see it, this just proves that they don't do |
58 |
>>> their job. |
59 |
>>> |
60 |
>> |
61 |
>> Because quizes lacks some questions? You see - we should not shorten |
62 |
>> them, but on the contrary - increase the question's number. |
63 |
>> |
64 |
>> So, that's a not good argument. Quizes may be flawed, but not in the way |
65 |
>> of 'over complicated', definitely. |
66 |
> |
67 |
> Sure. If the quizzes take so much nobody will bother writing them, |
68 |
> there will be no more bad developers in Gentoo! |
69 |
|
70 |
|
71 |
Let's throw the quizes away and give everyone commit access. |
72 |
At least developers count will grow, for sure! |
73 |
|
74 |
No matter that distro would be totally ruined, your goal would be achieved. |
75 |
|
76 |
>>>>> Add an extra contribution period in which the candidate commits to |
77 |
>>>> the tree through Pull Requests. |
78 |
>>>> |
79 |
>>>> All mentors should track their recruits during first month. If mentor |
80 |
>>>> doesn't do this, this is "bad" mentor, IMO. It looks like some mentors |
81 |
>>>> think *ok, my mentee passed the quizzes, he is a developer, i can to |
82 |
>>>> nothing* |
83 |
>>> |
84 |
>>> This happens *after* being recruited. What I'm talking, there should be |
85 |
>>> more focus on doing and judging actual contributions before all that |
86 |
>>> formal crap. |
87 |
>>> |
88 |
>> |
89 |
>> Yep. And, guess, who is responsible for that? Mentor. Tadaaam! :-) |
90 |
> |
91 |
> You still don't get my point. My point is: if a recruit has proven me |
92 |
> (and possibly some other developers) that he can use grep, sed, awk, |
93 |
> whatever just fine, why does he have to bother answering some stupid |
94 |
> quiz question? |
95 |
> |
96 |
|
97 |
In current system recruiters can not track progress of every single |
98 |
mentee, that's what mentors for. |
99 |
|
100 |
If you want to prove your recruit is doing fine with all stuff - try not |
101 |
to fight with quizes, but provide easy-to-use solution for viewing |
102 |
commits for recruiters, that can be answers for quizes itself. |
103 |
|
104 |
Can recruit use grep/sed/awk? That's the commit where he did this pretty |
105 |
fine. |
106 |
|
107 |
Does he know about live/snapshot ebuilds? That's commit where he used it. |
108 |
|
109 |
Gathering apropriate commit links can be work, that could be done both |
110 |
by mentor and recruit. |
111 |
|
112 |
And yes, that would be also very time-consuming, but recruiters would be |
113 |
glad for speeding up review sessions. |
114 |
|
115 |
-- |
116 |
Best regards, Sergey Popov |
117 |
Gentoo developer |
118 |
Gentoo Desktop Effects project lead |
119 |
Gentoo Quality Assurance project lead |
120 |
Gentoo Proxy maintainers project lead |