1 |
On Wednesday 14 April 2004 16:28, Koon wrote: |
2 |
> As a new Gentoo developer but with a little experience of management of |
3 |
> open projects, here is my opinion on the subject. |
4 |
> |
5 |
> |
6 |
> 1- On NFP |
7 |
> --------- |
8 |
> |
9 |
> I was still recently a member of Gentoo user base, and I think what the |
10 |
> user base wants is be sure Gentoo will be there tomorrow and will be |
11 |
> free (as in beer and as in freedom). The average user probably also |
12 |
> wants to have more leverage on the direction it's heading. The developer |
13 |
> wants to make sure his work will not be stolen by a dark corporate |
14 |
> conspiracy and will remain free. |
15 |
> |
16 |
> My experience shows that true democracy in open projects is not viable. |
17 |
> When truly implemented, it's an illusion or/and a innovation killer. |
18 |
> Resources are limited, choices must be made, conflicts must be resolved, |
19 |
> someone must have final word. Management is a vertical thing, not an |
20 |
> horizontal one. There must be a benevolent dictator or a small group of |
21 |
> managers. That doesn't seem right, but it's the only effective way. What |
22 |
> if the manager(s) does not follow the will of the community ? What if he |
23 |
> becomes a bad dictator, oppressing his people ? The open source |
24 |
> ecosystem has its answer : the project can be forked. The current lead |
25 |
> should do his best to avoid that, change his views, step down and let |
26 |
> another lead take over. But there always is this ultimate solution. It's |
27 |
> a painful process but ultimately the community will choose. They will |
28 |
> vote with their feet. |
29 |
> |
30 |
> So I think the best is a closed model. Since it's more or less the way |
31 |
> it works for now, I think the community can accept it. As long as the |
32 |
> the main goal ('forever free') is clearly built-in. |
33 |
> |
34 |
> A lot of devs are here because they love Gentoo's way : technical |
35 |
> issues, not political issues. A lot of devs will leave (and are leaving) |
36 |
> if political issues take over. It's the midlife management crisis for |
37 |
> the Gentoo project, we won't go through unaffected. But hopefully we |
38 |
> will go through stronger and more effective. |
39 |
|
40 |
One of the crises of growth, take any organization book and they are in there. |
41 |
But indeed hopefully we will grow stronger and find our own, gentoo, |
42 |
solution. |
43 |
|
44 |
> |
45 |
> Gentoo is a lot of things. To ensure that the open source ecosystem can |
46 |
> be applied to it, we must consider them separately. Each could have its |
47 |
> own lead and fork possibility : |
48 |
> |
49 |
> - Gentoo is portage |
50 |
> The portage technology is the core of the Gentoo system. It's difficult |
51 |
> to change without changing the ebuild tree, but could be changed. |
52 |
> |
53 |
> - Gentoo is a tree of packages using portage |
54 |
> The official portage tree with its arches, ebuilds, stable keywords. |
55 |
> Alternative/additional portage trees can exist. |
56 |
> |
57 |
> - Gentoo is a distribution using the portage tree |
58 |
> A distribution is a little more than a package tree : it has releases, |
59 |
> security updates, installation ISOs, a mirror network... Closely related |
60 |
> to the precedent, but could be separated from it. |
61 |
> |
62 |
> - Gentoo is a helpful community using the distribution |
63 |
> The forums and the mailing-lists are also what makes Gentoo a success. A |
64 |
> lot of users of others distributions find the Gentoo forums more useful |
65 |
> than their own dist forums. A fork at community level is probably not |
66 |
> possible. |
67 |
> |
68 |
> Should a single NFP cover the whole thing ? Or should you have a portage |
69 |
> open source project, a tree+distribution NFP, and a community with its |
70 |
> own hierarchy of moderators ? |
71 |
|
72 |
For me, gentoo is a philosophy. It is a way of handling problems, it is a way |
73 |
of having optimal choice. For me the combination of the features, including |
74 |
the openness of the comunity makes gentoo what it is. This gave me the desire |
75 |
to be part of it. As such I don't think that they can be separated. |
76 |
|
77 |
> |
78 |
> |
79 |
> 3- on Coop |
80 |
> ---------- |
81 |
> |
82 |
> I think the coop idea is very interesting and innovative. But I also |
83 |
> think it can easily be separated from the NFP/Management issues. The |
84 |
> coop(s) decide where money is spent. The university-driven coop(s) can |
85 |
> fund a particular developer if they want his particular work to advance |
86 |
> full-time. The coop(s) don't have to have the same lead as the NFP(s). |
87 |
> If they don't like the way it goes, they can just cut the money flow and |
88 |
> induce a fork by funding a parallel project. Vote using their wallet. |
89 |
|
90 |
I like the idea too. It makes a lot of sense, but the coop should be a |
91 |
separate entity. It should have it's own goals and policies. While the coop |
92 |
("oss support foundation") would be an important source of income for the NFP |
93 |
"gentoo foundation" should be created as soon as possible as grinding out the |
94 |
goals and policies of the coop will probably still take quite a while. |
95 |
|
96 |
> |
97 |
> |
98 |
> Conclusion |
99 |
> ---------- |
100 |
> |
101 |
> I think we can have a global solution with a distribution (under one or |
102 |
> several NFP projects with closed leadership), a community not directly |
103 |
> depending on the distribution (that can choose with their feet between |
104 |
> forks) and separate coop(s) (that can influence where it goes by using |
105 |
> its money). I think I rephrase what klieber already said, but I like to |
106 |
> be verbose, despite my bad English :) |
107 |
|
108 |
I'm not a native English speaker, but I don't think your English is bad at |
109 |
all. Besides that, I actually agree with your points, they make a lot of |
110 |
sense. |
111 |
|
112 |
Paul |
113 |
|
114 |
-- |
115 |
Paul de Vrieze |
116 |
Gentoo Developer |
117 |
Mail: pauldv@g.o |
118 |
Homepage: http://www.devrieze.net |