1 |
On Fri, 2003-06-27 at 14:47, Jon Portnoy wrote: |
2 |
> On Fri, Jun 27, 2003 at 02:21:35PM -0500, Matthew Kennedy wrote: |
3 |
> > I personally feel that it may be a good time to reconsider making |
4 |
> > gentoo-core a publicly read-only list. Several users I know (some |
5 |
> > personally) are irate that we appear to be a "behind-closed-doors" |
6 |
> > project. Yes, this stems from the recent fork announcement. |
7 |
> > |
8 |
> > gentoo-core is where we do our planning, discuss management structure, |
9 |
> > discuss technical questions (of which we already try to CC gentoo-dev |
10 |
> > out of courtesy) and architecture issues. |
11 |
> > |
12 |
> > I hope that we can open -core as a publicly read-only list to involve |
13 |
> > our community more. We owe this much to our user-base in my opinion. |
14 |
> > For these reasons, I have CC'd this to -dev. |
15 |
> > |
16 |
> > Matt |
17 |
> > |
18 |
> |
19 |
> We've already gone over the reasons -core should remain private |
20 |
> (discussing security before it goes public and personal information). I |
21 |
> think that a better approach would be to enforce a policy of all |
22 |
> technical stuff and anything that isn't sensitive being discussed on |
23 |
> -dev rather than -core. |
24 |
|
25 |
Moving all development talk to the, well, -dev list would be the best |
26 |
solution. |
27 |
|
28 |
Correct me if I'm wrong, but I have heard four main reasons to keep |
29 |
-core private: |
30 |
|
31 |
1) Gentoo should speak with a unified voice, and by keeping the |
32 |
community excluded from -core, announcements and policy changes can be |
33 |
discussed without a public display of fragmentation. |
34 |
|
35 |
2) Devs make mistakes, and would rather "fall on their face" in private |
36 |
rather than in public. |
37 |
|
38 |
3) -core would quickly become cluttered with mis-postings. |
39 |
|
40 |
4) -core is boring anyway, why would you want to read it? |
41 |
|
42 |
While these four reasons, as well as the security issues mentioned |
43 |
above, look to be good ones on the surface, they become marginal when |
44 |
examined and all serve the purpose of setting the developers apart and |
45 |
away from their community. |
46 |
|
47 |
Security is not a valid excuse to keep things secret - Bugtraq and its |
48 |
kind already bring things out in the open quickly enough. That dead |
49 |
horse has been beaten enough. |
50 |
|
51 |
Gentoo can speak with a unified voice while still allowing discussion. |
52 |
For example, it seems like once a month someone posts to -dev offering |
53 |
to re-implement portage or something like it in another language - Java, |
54 |
C, C++, Perl, whatever. The standard answer from a dev is invariably |
55 |
"We've discussed this on -core before and decided it wasn't a good |
56 |
idea." Now, the user who offered of himself to make what he saw as an |
57 |
improvement feels shut out and unwanted. Had there existed -core |
58 |
archive, the user could have read through the reasons and understood why |
59 |
- things are not always as obvious as they seem. Even better, the user |
60 |
might have looked through the archives and discovered the said |
61 |
discussions before posting, saving everyone a headache. Gentoo has |
62 |
stayed with python despite many offers - what harm does a valid, public |
63 |
evaluation do? |
64 |
|
65 |
Devs do make mistakes - Daniel Robbins recently slipped on some |
66 |
unmasking and there were a lot of questions in -user and on the boards. |
67 |
Obviously, no one wants to make this kind of mistake, but the end result |
68 |
wasn't bad - it was corrected quickly, and no one thinks Daniel is |
69 |
incompetent. Everyone makes mistakes. |
70 |
|
71 |
If -core is read-only, there won't be mis-postings to it by unwary |
72 |
users. -core might be very boring, but even most Gentoo users' eyes |
73 |
would glaze over trying to parse the kernel code. Does that mean the |
74 |
kernel is distributed as a binary, because 'it would be boring' to try |
75 |
and improve the source code for those few that would want to? Perhaps |
76 |
the devs might have an idea that a non-"official"-dev can implement |
77 |
quickly; everyone benefits. |
78 |
|
79 |
Currently, -dev isn't a developer list; most of the e-mail is users' |
80 |
suggestions discussed by other users and the occasional dev, as well as |
81 |
the occasional mis-post targeted for -user. Perhaps in the future, if |
82 |
-core would be accessible to the community that drive it, -dev could be |
83 |
the buffer zone between the final work of the developers and the user |
84 |
community. Right now, there's a feeling that the developers are |
85 |
shutting themselves in an ivory tower. That's not good for community. |
86 |
Gentoo's social contract has always said it will not "hide its |
87 |
problems", but has continued to keep its core development decisions |
88 |
closed. |