Therefore in free software projects this law could be that 20% of committers produce 80% of commits.
Let's go to analyse some projects to do a first approximation checking of the law, the list of chosen projects:
- LibreOffice
- CloudStack
- Evince
- CVSAnalyt
The analysis assesses number of commits and committers in the 2013 and aggregate committers to 20% of commits, and calculate percentage in number of committers. therefore the results are
- CloudStack
- Linux
Therefore, the Pareto's Law does not seem true in these big projects (Maybe we should define accurately when a committer is really a contributor, a contribuitor with 1 commit is a contributor?.)
But finally, this law is valid for small project, try to check the others, for instance, evince (a pdf visor) and even CVSAnalyt.
- Evince
- CVSAnalyt
In these projects, the law does not seem to be true neither.
References:
[1] Metrics Grimoire - http://metricsgrimoire.github.io/
[2] CVSAnalY - https://github.com/MetricsGrimoire/CVSAnalY
[3] LibreOffice repository - git://anongit.freedesktop.org/libreoffice/core
[4] CloudStack repository - https://git-wip-us.apache.org/repos/asf/cloudstack.git
[5] Linux repository - https://github.com/torvalds/linux
Note:
Using query in CVSAnalY to get data
SELECT COUNT(distinct(scmlog.id)) as total, people.email FROM scmlog, actions, people WHERE scmlog.id = actions.commit_id AND scmlog.author_id = people.id AND YEAR(scmlog.date) = 2013 GROUP BY people.email ORDER BY total DESC INTO OUTFILE '/tmp/data.csv' FIELDS TERMINATED BY ',' ENCLOSED BY '"' LINES TERMINATED BY '\n';
No hay comentarios:
Publicar un comentario