17

How can I get statistics about how much (in percents) the Linux kernel source code changes in one year?

Mat
  • 52,586
LanceBaynes
  • 40,135
  • 97
  • 255
  • 351
  • The Linux Foundation made a study on "Who writes Linux?" in 2009 that might be interesting, too. – sr_ Jan 27 '12 at 14:00
  • 2
    I don't know what sort of question you are looking to answer, but looking at lines of code as a proxy measure for anything other than lines of code is almost certain to be misleading. For example, an experimental scheduler can be added that almost no one uses except for experiment. Is that 2000 lines of change or 0? – msw Jan 27 '12 at 14:09
  • What do you want to prove with that statistic? Never trust a statistic that you did not falsify yourself... – Nils Jan 28 '12 at 21:14

3 Answers3

18

Your best bet would be to get statistical data about the commits themselves. gitstats is designed for statistical analysis of Git repositories (which is the VCS that Linux uses), and should fit your needs.

Chris Down
  • 125,559
  • 25
  • 270
  • 266
9

You'll find this kind of information on lwn.net website. It's made by Jonathan Corbet and has an in-depth coverage of Linux kernel. Jonathan Corbet is a highly-recognized contributor to the Linux kernel community.

On the Kernel Index page, you can look at "who wrote ..." pages. They contains detailled information about changes in linux kernel from a version to another.

Linux foundation has gathered this information into an official document. There's also a 2010 retrospective made by Greg Kroah-Hartman.

Edit: found an official source.

Coren
  • 5,010
9

What you are looking for can be found on the Ohloh website, which by the way indexes the Linux GIT repository. There you will see a graph showing you how much the kernel has changed over 1 yr, 3 yrs, 5 yrs, 10 yrs or All. By default it will show you the statistics for the source code but you can also get statistics about Languages, Committers, Commits. You can then manually calculate the change %. The change in source code between 2010 and 2011 is up 11.4%.

Ura
  • 534
  • 2
    what is that in 2005? why was there a very-very big "boom"? – LanceBaynes Feb 05 '12 at 06:56
  • 5
    In 2005, Linus created GIT. The merging of previous branches into GIT reflects the spike. This is just an inconsistency in data while moving from one revision control system to another. – Ura Feb 06 '12 at 17:27