1 |
On 23/04/18 21:01, Robin H. Johnson wrote: |
2 |
> On Mon, Apr 23, 2018 at 05:24:41AM +1200, Kent Fredric wrote: |
3 |
>> On Sun, 15 Apr 2018 18:22:48 +0100 |
4 |
>> "M. J. Everitt" <m.j.everitt@×××.org> wrote: |
5 |
>> |
6 |
>>> -- if Dev does not set Devaway, and/or devaway period is over ~6months |
7 |
>>> (say) and activity has fallen to zero .. commit privs get automatically |
8 |
>>> revoked (by script, not by human). An automated email is sent out to |
9 |
>>> that Dev, encouraging them to contact [insert project here] (eg. |
10 |
>>> Council, ComReS, DevRel, etc) if there is good reason for the absence, |
11 |
>>> the privs can be reinstated after a petition has been received and reviewed. |
12 |
>> The problem here is that not all contribution is "visible" on any |
13 |
>> metrics system. |
14 |
>> |
15 |
>> This is only useful in regards to measuring *commit* contribution. |
16 |
>> |
17 |
>> But there are many other ways to contribute to Gentoo, and not all of |
18 |
>> them monitorable by automated tools. |
19 |
> In one of the threads leading up to this pre-GLEP, I proposed a |
20 |
> requiring testable metrics for all activity. It was dropped for |
21 |
> simplicity, as it's probably something that would merit it's own GLEP. |
22 |
> |
23 |
> It would NOT be a way to measure >1 contribution, but just a way to show |
24 |
> the date of last contribution for each project where a dev is active. |
25 |
> Specifically it would NOT disclose what that action was (because the |
26 |
> action in itself may be protected/private information). |
27 |
> |
28 |
> Commits: last commit date (on both public & private repos). |
29 |
> Bugzilla: last comment OR action date (even if the bug is private) |
30 |
> Forums: date of last post OR mod-action (even if the post or mod-action |
31 |
> is in a hidden/private forum) |
32 |
> |
33 |
> Bugzilla & Commits already do this effectively, it would be exposing the |
34 |
> data more readily, and adding similar functionality for other projects. |
35 |
> |
36 |
> Think like an IRC bot responding to '!seen', but it only tells you when |
37 |
> the person was seen [optionally for a specific project], and not exactly |
38 |
> what they were saying/doing. |
39 |
> |
40 |
I'm inclined to support the idea that <something> is marginally better |
41 |
than <nothing> and could be used as some form of evidence should there |
42 |
be need to substantiate a claim either way, and make it much less |
43 |
subjective ... |
44 |
|
45 |
Not looking for a perfect solution here, that IS impossible, just |
46 |
something that would aid metrics... |