| |||||||||||||
Gentoo debates recruitment schemesGentoo debates recruitment schemesPosted Jul 21, 2012 17:07 UTC (Sat) by tetromino (subscriber, #33846)In reply to: Gentoo debates recruitment schemes by akeane Parent article: Gentoo debates recruitment schemes
First, Gentoo has 280 people with commit rights to the main repository. Numerous experiments have shown that an average human brain can't deal with relationships in a group of >100 peers or so (the size of an ancient tribal band). So some degree of bureaucracy/HR is absolutely required for a project of such size, otherwise you get chaos. Second, it's not enough for the answers to quizzes to be discoverable. Much of it is information you need to know by heart, so you don't make mistakes and break user machines or community rules in the first place, instead of merely being able to google for how to fix the mess you had made after the fact.
(Log in to post comments)
Gentoo debates recruitment schemes Posted Jul 21, 2012 17:10 UTC (Sat) by tetromino (subscriber, #33846) [Link] 280 people with commit rights s/280/267/. Turns out I can't count.
Gentoo debates recruitment schemes Posted Jul 21, 2012 20:49 UTC (Sat) by boudewijn (subscriber, #14185) [Link]
KDE currently has 2626 commit accounts -- http://websvn.kde.org/trunk/kde-common/accounts?revision=... -- and no quizzes. Of course, not all of those are active right now, and there's plenty chaos, but a remarkable amount of productivity and very little disruptive chaos.
Gentoo debates recruitment schemes Posted Jul 26, 2012 19:17 UTC (Thu) by jospoortvliet (subscriber, #33164) [Link]
yup, and openSUSE also manages to get by just fine without such bureaucracies. I won't call it ridiculous outright but it certainly doesn't sit well by me.
Gentoo debates recruitment schemes Posted Aug 2, 2012 12:22 UTC (Thu) by TRauMa (guest, #16483) [Link]
But there is a company behind it, which means full time employees and a hidden bureaucracy layer. Gentoo doesn't have that.
Gentoo debates recruitment schemes Posted Sep 6, 2012 9:34 UTC (Thu) by jospoortvliet (subscriber, #33164) [Link]
True, SUSE and the other companies behind openSUSE mudd the waters a bit here. But Gentoo has corporate involvement too, does it not?
Gentoo debates recruitment schemes Posted Oct 3, 2012 23:51 UTC (Wed) by Duncan (guest, #6647) [Link]
> SUSE ... openSUSE.
> Gentoo has corporate involvement too, does it not?
Not in the same way. While various companies (corporation and otherwise) may provide mirrors and/or donate hardware and the like, and google itself is a notable example of a gentoo downstream (with chromeOS), gentoo seems to be much like the kde of the front-page article, in that it's very much individual developer driven, with companies as such not allowed to get too close.
I believe there's a parallel to debian there. Both gentoo and debian are 100% community distros without a sponsoring corporate parent, as such. The model is quite different from that of fedora/rh, or opensuse/suse, or mandriva (closer to mageia I guess).
It occurs to me that I don't know where arch fits in, community/corporate-sponsorship-wise.
Duncan
Gentoo debates recruitment schemes Posted Jul 22, 2012 11:01 UTC (Sun) by mchandras (subscriber, #81428) [Link]
This number is a bit inaccurate. We currently have 41 open retirement bugs which means that the potential number of active contributors is 227. Plus, we do have some people listed in that list that have no commit access so, again, the number of potential active contributors with commit access is ~220. Then again, not all of them are active. I am quite confident that there are no more than 190 active developers right now in Gentoo. This is not as bad as it may sound, since this number remains constant for the last 3-4 years.
Gentoo debates recruitment schemes Posted Jul 22, 2012 9:16 UTC (Sun) by akeane (subscriber, #85436) [Link]
>First, Gentoo has 280 people with commit rights to the main repository. Numerous experiments have shown that an average human brain can't deal with relationships in a group of >100 peers or so (the size of an ancient tribal band). So some degree of bureaucracy/HR is absolutely required for a project of such size, otherwise you get chaos.
That's an interesting point, but isn't at least some of the benefits of using a software repository/version control is that some of this complexity (i.e. multiple folks changing the same codebase) is contained?
>Much of it is information you need to know by heart, so you don't make mistakes and break user machines or community rules in the first place, instead of merely being able to google for how to fix the mess you had made after the fact.
I meant using google to pass the quiz ;-)
Also it really possible/useful even these days, to know everything by heart? After many moons of coding I certainly don't bother to remember everything I do or have done if I can find it easily. Also having to fix things last minute via muscle memory ESC ESC wq: :-) implies a rushed deadline, and/or bad QA and testing cycle prior to release...
I guess fundamentally I find it difficult to believe a quiz is a better way to assess somebody, rather than previous interactions with maintainers say, or a community reputation - which can't be "faked" with l33t google skilz.
Anyway, thanks for taking the time to answer my points, it's certainly interesting to see how different parts of the community are trying to handle these problems.
Gentoo debates recruitment schemes Posted Jul 23, 2012 15:41 UTC (Mon) by sumanah (guest, #59891) [Link] Disclaimer: I've never taken the Gentoo quizzes, and I'm talking more generally about the merits and disadvantages of bureaucratic processes.I guess fundamentally I find it difficult to believe a quiz is a better way to assess somebody, rather than previous interactions with maintainers say, or a community reputation - which can't be "faked" with l33t google skilz. One thing to consider: the merits of objective procedures. When a community says "here's the procedure for how to get such-and-such privileges," and they are clear and scalable and don't depend on forming relationships with specific people, then that community reassures its members that it's being transparent, objective, and fair in distributing those privileges. This reduces a risk of unfair discrimination based on prejudice, and makes the user interface of the community a lot clearer for newbies. Of course, the procedures themselves become a center of discussion around what actually constitutes merit and what skills should lead to which privileges, and that sort of discussion should be happening anyway. And sometimes communities struggle to consistently apply the rules, but at least they're talking about that instead of letting the issue simmer beneath the surface. Bureaucracies scale better than personal judgments, and can do better at fairness and accessibility. But they also put off some kinds of participants, and, if improperly administered, can cause inefficiencies and frustration. I applaud the Gentoo community for thinking and working on this from both approaches; I agree with Nathan Willis's summary: "the specifics of the training process are less important than the fact that is it deliberate and guided by active mentors and recruiters".
Gentoo debates recruitment schemes Posted Jul 24, 2012 18:49 UTC (Tue) by akeane (subscriber, #85436) [Link]
>One thing to consider: the merits of objective procedures. When a community says "here's the procedure for how to get such-and-such privileges," and they are clear and scalable and don't depend on forming relationships with specific people, then that community reassures its members that it's being transparent, objective, and fair in distributing those privileges. This reduces a risk of unfair discrimination based on prejudice, and makes the user interface of the community a lot clearer for newbies. Of course, the procedures themselves become a center of discussion around what actually constitutes merit and what skills should lead to which privileges, and that sort of discussion should be happening anyway. And sometimes communities struggle to consistently apply the rules, but at least they're talking about that instead of letting the issue simmer beneath the surface.
google: how do I get the procedure for how to get such-and-such privileges?
So, google in a clear and scalable, objective way answers that, how do you tell you tell it's not google doing that, guess it's the policy vs mechanism thing...
You may think these things need to be scalable, more bureaucratic, but I notice that there are many linux devices physically around me right now, like the embedded MIPS linux controller in my TV, or ARM linux in my wireless router, phone...
And this is something that got created without a huge bureaucracy, you can just turn up on lkml without a quiz and see how it goes.
At the end of the day, none of these distributions would exist without the kernel that somehow got produced (with it's millions of lines of code and thousands of contributors) without quizzes/HR stuff.
I guess, I am asking why not just adopt that model, as it obviously works, seems scalable, produced good stuff over the last two decades?
Or are the people who actually produce the stuff that the distros use/sell/possibly make money with, on a daily basis just a bunch of statistics to be processed/filtered according to, well, a bureaucratic process?
Play Knee Deep in zDoom :-) and stop trying to spell bureaucracy
Gentoo debates recruitment schemes Posted Jul 24, 2012 21:06 UTC (Tue) by raven667 (subscriber, #5198) [Link]
The kernel model is to have a Benevolent Dictator for Life at the top who can act however they want. You can show up on LKML and drop some patches and see if you get quizzed or not...Linux code review is famously harsh and sometimes the procedures to get changes accepted are arbitrary and inscrutable. You notice that vendors shipping a lot of linux derivatives but just because you shipped millions of units doesn't mean that your changes will get added to the Linus kernel on kernel.org, just ask the Android team about that. 8-)
Just because Linux development works doesn't mean that it is the best method or that it's methods are applicable to other organizations.
Gentoo debates recruitment schemes Posted Aug 6, 2012 9:59 UTC (Mon) by philomath (guest, #84172) [Link]
>Numerous experiments have shown that an average human brain can't deal with relationships in a group of >100 peers or so
Interesting.
Gentoo debates recruitment schemes Posted Aug 6, 2012 11:28 UTC (Mon) by jwakely (subscriber, #60262) [Link]
|
|||||||||||||