productivity of mediawiki developers

classic Classic list List threaded Threaded
15 messages Options
Reply | Threaded
Open this post in threaded view
|

productivity of mediawiki developers

rupert THURNER-2
hi,

is there a statistics about mediawiki developer productivity? i just
fell over a couple of pages and i am quite impressed i must say:
* gabriel, https://github.com/gwicke, 2'300 commits last year
* jeroen, https://github.com/JeroenDeDauw, 3'700 commits a year
* ori, https://github.com/atdt, 1'700
* james, https://github.com/jdforrester, 1'200
* yuri, https://github.com/nyurik, 900
* matt: https://github.com/mattflaschen, 400 commits

rupert

_______________________________________________
Wikitech-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: productivity of mediawiki developers

Pine W
Perhaps some Working Wikipedian and Tireless Contributor barnstars are
appropriate for that group!

Pine

On Sat, Apr 2, 2016 at 12:06 PM, rupert THURNER <[hidden email]>
wrote:

> hi,
>
> is there a statistics about mediawiki developer productivity? i just
> fell over a couple of pages and i am quite impressed i must say:
> * gabriel, https://github.com/gwicke, 2'300 commits last year
> * jeroen, https://github.com/JeroenDeDauw, 3'700 commits a year
> * ori, https://github.com/atdt, 1'700
> * james, https://github.com/jdforrester, 1'200
> * yuri, https://github.com/nyurik, 900
> * matt: https://github.com/mattflaschen, 400 commits
>
> rupert
>
> _______________________________________________
> Wikitech-l mailing list
> [hidden email]
> https://lists.wikimedia.org/mailman/listinfo/wikitech-l
_______________________________________________
Wikitech-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: productivity of mediawiki developers

Florian Schmidt
In reply to this post by rupert THURNER-2
Hi rupert,

I'm pretty sure http://korma.wmflabs.org/browser/scm.html and other pages of korma will give you some information. But keep in mind, that productivity doesn't only mean the number of commits, or the line of codes :)

Best,
Florian

-----Ursprüngliche Nachricht-----
Von: Wikitech-l [mailto:[hidden email]] Im Auftrag von rupert THURNER
Gesendet: Samstag, 2. April 2016 21:06
An: Wikimedia developers <[hidden email]>
Betreff: [Wikitech-l] productivity of mediawiki developers

hi,

is there a statistics about mediawiki developer productivity? i just fell over a couple of pages and i am quite impressed i must say:
* gabriel, https://github.com/gwicke, 2'300 commits last year
* jeroen, https://github.com/JeroenDeDauw, 3'700 commits a year
* ori, https://github.com/atdt, 1'700
* james, https://github.com/jdforrester, 1'200
* yuri, https://github.com/nyurik, 900
* matt: https://github.com/mattflaschen, 400 commits

rupert

_______________________________________________
Wikitech-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wikitech-l


_______________________________________________
Wikitech-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: productivity of mediawiki developers

Pine W
Yeah. It would be interesting to have multiple measures of "productivity"
for technical contributors, including code review.

Quim: is this something that's within Technical Collaboration's scope? If
not, perhaps it's something that I could work on developing later this year.

Pine

On Sat, Apr 2, 2016 at 12:10 PM, Florian Schmidt <
[hidden email]> wrote:

> Hi rupert,
>
> I'm pretty sure http://korma.wmflabs.org/browser/scm.html and other pages
> of korma will give you some information. But keep in mind, that
> productivity doesn't only mean the number of commits, or the line of codes
> :)
>
> Best,
> Florian
>
> -----Ursprüngliche Nachricht-----
> Von: Wikitech-l [mailto:[hidden email]] Im
> Auftrag von rupert THURNER
> Gesendet: Samstag, 2. April 2016 21:06
> An: Wikimedia developers <[hidden email]>
> Betreff: [Wikitech-l] productivity of mediawiki developers
>
> hi,
>
> is there a statistics about mediawiki developer productivity? i just fell
> over a couple of pages and i am quite impressed i must say:
> * gabriel, https://github.com/gwicke, 2'300 commits last year
> * jeroen, https://github.com/JeroenDeDauw, 3'700 commits a year
> * ori, https://github.com/atdt, 1'700
> * james, https://github.com/jdforrester, 1'200
> * yuri, https://github.com/nyurik, 900
> * matt: https://github.com/mattflaschen, 400 commits
>
> rupert
>
> _______________________________________________
> Wikitech-l mailing list
> [hidden email]
> https://lists.wikimedia.org/mailman/listinfo/wikitech-l
>
>
> _______________________________________________
> Wikitech-l mailing list
> [hidden email]
> https://lists.wikimedia.org/mailman/listinfo/wikitech-l
>
_______________________________________________
Wikitech-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: productivity of mediawiki developers

Federico Leva (Nemo)
In reply to this post by rupert THURNER-2
For statistics, see
https://www.mediawiki.org/wiki/Development_statistics . As a starting
point to look into what you call "productivity", I usually use:
* https://www.openhub.net/orgs/wikimedia
* http://koti.kapsi.fi/~federico/crstats/core.txt +
http://koti.kapsi.fi/~federico/crstats/extensions.txt

Nemo

_______________________________________________
Wikitech-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: productivity of mediawiki developers

Joaquin Oltra Hernandez
Lines added: Last 365 days: 20,176,017
Lines removed: Last 365 days: 14,378,469

Is this ^ right? Those are really big numbers!

On Sun, Apr 3, 2016 at 10:39 PM, Federico Leva (Nemo) <[hidden email]>
wrote:

> For statistics, see https://www.mediawiki.org/wiki/Development_statistics
> . As a starting point to look into what you call "productivity", I usually
> use:
> * https://www.openhub.net/orgs/wikimedia
> * http://koti.kapsi.fi/~federico/crstats/core.txt +
> http://koti.kapsi.fi/~federico/crstats/extensions.txt
>
> Nemo
>
>
> _______________________________________________
> Wikitech-l mailing list
> [hidden email]
> https://lists.wikimedia.org/mailman/listinfo/wikitech-l
>
_______________________________________________
Wikitech-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: productivity of mediawiki developers

Quim Gil-2
In reply to this post by Pine W
On Sat, Apr 2, 2016 at 10:13 PM, Pine W <[hidden email]> wrote:

> Yeah. It would be interesting to have multiple measures of "productivity"
> for technical contributors, including code review.
>
> Quim: is this something that's within Technical Collaboration's scope? If
> not, perhaps it's something that I could work on developing later this year.
>

The first question to answer is what information are you looking for when
you want to measure developers' "productivity". What would be the
motivation of that estimation? What is the motivation behind this thread?

--
Quim Gil
Engineering Community Manager @ Wikimedia Foundation
http://www.mediawiki.org/wiki/User:Qgil
_______________________________________________
Wikitech-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: productivity of mediawiki developers

Niklas Laxström
2016-04-04 17:02 GMT+03:00 Quim Gil <[hidden email]>:
> The first question to answer is what information are you looking for when
> you want to measure developers' "productivity". What would be the
> motivation of that estimation? What is the motivation behind this thread?

One reason comes to me mind. My gut feeling is that we are not very
good at consistently giving recognition for technical work. One
possible reason is that we do not have clear and understandable
metrics or promote those metrics enough. Nor am I aware of any process
for awards and celebration (The Academy Awards would be an example in
another context, also Wikipedian of the year).

As an example, I recall vaguely that during the Bugzilla times we used
to have regular emails on wikitech-l with list of people who closed
most bugs.

Having some metrics for different activities could stir up some
healthy competition (also unhealthy if we are not careful) and of
course there is a lot of important work that is not visible from the
numbers only.

I am not expert on this subject, but I think developers (especially
volunteers, but also others) are more likely to stick around if they
feel that their work is recognized and appreciated. For the latter we
already know that we should improve our code review process.

  -Niklas

_______________________________________________
Wikitech-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: productivity of mediawiki developers

Joaquin Oltra Hernandez
I agree we should give recognition and encouragement to devs, but I think
there are other ways to do it we could think about besides sheer number of
commits, +2s or lines modified.

I personally think that rewarding high numbers encourages quantity over
quality (only big numbers are recognized) and also encourages a culture of
hero developers[1] that is discouraging for casual or new volunteers and
grows our silos bigger.

I, for example, value better good -1 code reviews (well written,
thoughtful, and when you learn new things), rather than 10s of +2s, and for
example IMO a reviewer price would be hand picked from nominations for this
kind of reviews rather than automatically picked from top number of +2s.

[1]: There's always a small group of *heroes* that are highly productive
because -besides them being great developers- of the background knowledge,
familiarity with the code bases and other developers that are always going
to be in the top of most these metrics.
On Apr 4, 2016 5:22 PM, "Niklas Laxström" <[hidden email]> wrote:

> 2016-04-04 17:02 GMT+03:00 Quim Gil <[hidden email]>:
> > The first question to answer is what information are you looking for when
> > you want to measure developers' "productivity". What would be the
> > motivation of that estimation? What is the motivation behind this thread?
>
> One reason comes to me mind. My gut feeling is that we are not very
> good at consistently giving recognition for technical work. One
> possible reason is that we do not have clear and understandable
> metrics or promote those metrics enough. Nor am I aware of any process
> for awards and celebration (The Academy Awards would be an example in
> another context, also Wikipedian of the year).
>
> As an example, I recall vaguely that during the Bugzilla times we used
> to have regular emails on wikitech-l with list of people who closed
> most bugs.
>
> Having some metrics for different activities could stir up some
> healthy competition (also unhealthy if we are not careful) and of
> course there is a lot of important work that is not visible from the
> numbers only.
>
> I am not expert on this subject, but I think developers (especially
> volunteers, but also others) are more likely to stick around if they
> feel that their work is recognized and appreciated. For the latter we
> already know that we should improve our code review process.
>
>   -Niklas
>
> _______________________________________________
> Wikitech-l mailing list
> [hidden email]
> https://lists.wikimedia.org/mailman/listinfo/wikitech-l
_______________________________________________
Wikitech-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: productivity of mediawiki developers

Pine W
In reply to this post by Niklas Laxström
Niklas puts it well. Analogously, in sports like baseball there are lots of
statistics about players, coaches, teams, divisions, and leagues. Awards
are given based strictly on quantities, as well as more subjectively on
qualities for recognitions such as Rookie of the Year and Most Valuable
Player.

Wikimedia technical development is a team sport, and that unlike on the
content side of Wikipedia where there can be rival views, I think that on
technical matters almost everyone collaborates toward compatible goals.

I've been thinking about how to ensure that *quality* is valued alongside
*quantity*. We struggle do this balance well in the US health care system
when we evaluate hospitals and doctors, and WMF stuggles to do this well
when the Community Resources and Evaluation teams evaluate grant proposals
and the performance of Wikimedia affiliates. I'm very interested in ideas
about how to estimate the quality of contributions (including code review!)
as well as the quantity of contributions.

Pine
On Apr 4, 2016 08:22, "Niklas Laxström" <[hidden email]> wrote:

> 2016-04-04 17:02 GMT+03:00 Quim Gil <[hidden email]>:
> > The first question to answer is what information are you looking for when
> > you want to measure developers' "productivity". What would be the
> > motivation of that estimation? What is the motivation behind this thread?
>
> One reason comes to me mind. My gut feeling is that we are not very
> good at consistently giving recognition for technical work. One
> possible reason is that we do not have clear and understandable
> metrics or promote those metrics enough. Nor am I aware of any process
> for awards and celebration (The Academy Awards would be an example in
> another context, also Wikipedian of the year).
>
> As an example, I recall vaguely that during the Bugzilla times we used
> to have regular emails on wikitech-l with list of people who closed
> most bugs.
>
> Having some metrics for different activities could stir up some
> healthy competition (also unhealthy if we are not careful) and of
> course there is a lot of important work that is not visible from the
> numbers only.
>
> I am not expert on this subject, but I think developers (especially
> volunteers, but also others) are more likely to stick around if they
> feel that their work is recognized and appreciated. For the latter we
> already know that we should improve our code review process.
>
>   -Niklas
>
> _______________________________________________
> Wikitech-l mailing list
> [hidden email]
> https://lists.wikimedia.org/mailman/listinfo/wikitech-l
_______________________________________________
Wikitech-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: productivity of mediawiki developers

Federico Leva (Nemo)
In reply to this post by Joaquin Oltra Hernandez
 > I, for example, value better good -1 code reviews

Same here. There are statistics for -1 too, two clicks away from the
link I provided in the previous message.
https://www.mediawiki.org/wiki/Gerrit/Reports/Code_review_activity

Nemo

_______________________________________________
Wikitech-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: productivity of mediawiki developers

Jon Robson-2
I think a review is valuable no matter what the score... with the
possible exception of -2 which I fear is probably a bit too aggressive
and unnecessary in our ecosystem for which reason the reading web team
agreed to avoid the use of -2 except to stop merges in progress that
were not ready. I review with score 0 quite a lot for instance.


On Tue, Apr 5, 2016 at 3:59 PM, Federico Leva (Nemo) <[hidden email]> wrote:

>> I, for example, value better good -1 code reviews
>
> Same here. There are statistics for -1 too, two clicks away from the link I
> provided in the previous message.
> https://www.mediawiki.org/wiki/Gerrit/Reports/Code_review_activity
>
> Nemo
>
>
> _______________________________________________
> Wikitech-l mailing list
> [hidden email]
> https://lists.wikimedia.org/mailman/listinfo/wikitech-l

_______________________________________________
Wikitech-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: productivity of mediawiki developers

Daniel Kinzler
I personally reserve -2 for "this is a fundamentally bad idea" or "this requires
community consensus before being implemented". Anything that is fixable in the
code should get a -1 or 0.

Btw, I personally prefer to get -1 reviews over 0 reviews, simply because it's
easier to spot them as "todo" on the gerrit dashboard. If gerrit would highlight
"stuff with new comments" more prominently, I'd probably use 0 more often.

Am 05.04.2016 um 16:18 schrieb Jon Robson:

> I think a review is valuable no matter what the score... with the
> possible exception of -2 which I fear is probably a bit too aggressive
> and unnecessary in our ecosystem for which reason the reading web team
> agreed to avoid the use of -2 except to stop merges in progress that
> were not ready. I review with score 0 quite a lot for instance.
>
>
> On Tue, Apr 5, 2016 at 3:59 PM, Federico Leva (Nemo) <[hidden email]> wrote:
>>> I, for example, value better good -1 code reviews
>>
>> Same here. There are statistics for -1 too, two clicks away from the link I
>> provided in the previous message.
>> https://www.mediawiki.org/wiki/Gerrit/Reports/Code_review_activity
>>
>> Nemo
>>
>>
>> _______________________________________________
>> Wikitech-l mailing list
>> [hidden email]
>> https://lists.wikimedia.org/mailman/listinfo/wikitech-l
>
> _______________________________________________
> Wikitech-l mailing list
> [hidden email]
> https://lists.wikimedia.org/mailman/listinfo/wikitech-l
>


_______________________________________________
Wikitech-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: productivity of mediawiki developers

Stas Malyshev
Hi!

> I personally reserve -2 for "this is a fundamentally bad idea" or "this requires
> community consensus before being implemented". Anything that is fixable in the
> code should get a -1 or 0.
>
> Btw, I personally prefer to get -1 reviews over 0 reviews, simply because it's
> easier to spot them as "todo" on the gerrit dashboard. If gerrit would highlight
> "stuff with new comments" more prominently, I'd probably use 0 more often.

I treat -1 as "this needs to be fixed before it can go in, but once it
is fixed it's good". Agree on -2. I use 0 for just commenting on things
where I do not feel qualified or entitled to review things but still
have something to say, like additional todo items or general discussion.
So, most reviews should be +1/-1.
--
Stas Malyshev
[hidden email]

_______________________________________________
Wikitech-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: productivity of mediawiki developers

Andre Klapper-2
In reply to this post by Florian Schmidt
On Sat, 2016-04-02 at 21:10 +0200, Florian Schmidt wrote:
> I'm pretty sure http://korma.wmflabs.org/browser/scm.html and other
> pages of korma will give you some information.

For the records, in the meantime korma.wmflabs.org got replaced by
https://wikimedia.biterg.io/ . Documentation is available at
https://www.mediawiki.org/wiki/Community_metrics#wikimedia.biterg.io

Cheers,
andre
--
Andre Klapper | Wikimedia Bugwrangler
http://blogs.gnome.org/aklapper/

_______________________________________________
Wikitech-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wikitech-l