1 |
On Fri, Feb 20, 2015 at 4:05 PM, hasufell <hasufell@g.o> wrote: |
2 |
> Rich Freeman: |
3 |
>> |
4 |
>> I like Robin's "kickstarter" suggestion in general as a potential way |
5 |
>> to improve things. I'll probably comment further in the other |
6 |
>> thread... |
7 |
>> |
8 |
> |
9 |
> I think the solution is to make less pressure on infra. That way we are |
10 |
> also less dependant on infra. |
11 |
> |
12 |
> This would be easily possible if gentoo was a collection of high-quality |
13 |
> overlays sharing the same base of policies and having tight |
14 |
> communication for inter-overlay matters, e.g. with a central bug tracker. |
15 |
> |
16 |
|
17 |
I certainly agree that we should try to focus on the elements of infra |
18 |
that are most essential. |
19 |
|
20 |
However, I'm not sure that your suggestion will necessarily reduce the |
21 |
impact on infra. Moving lots of stuff to overlays doesn't get rid of |
22 |
the need for a git repo unless we move EVERYTHING to overlays that |
23 |
aren't hosted by Gentoo. We might reduce the number of mirrors we |
24 |
need if we stop mirroring some packages, but we won't eliminate the |
25 |
need for mirrors unless we don't mirror anything. Etc. |
26 |
|
27 |
I imagine 95% of the burden of infra is in providing services at all. |
28 |
If we didn't host any mailing lists, that would reduce infra. If we |
29 |
just retired a few lists, that wouldn't really make much difference. |
30 |
|
31 |
Just something to keep in mind when trying to reduce our scope. I do |
32 |
generally think we should be thinking about what services are most |
33 |
important to host ourselves vs just utilizing stuff that is out there. |
34 |
|
35 |
-- |
36 |
Rich |