.

Tags:

At the palindrom date 11/11/11, WebKit project hit its monumental 100,000-th check in. Now, what happened from the moment it was born till this time? The following chart shows the source code revision (vertical) as the function of time (horizontal). Some icons are there to represent few products associated with WebKit, the position approximately resembles the era those products were made popular.

When I became a WebKit reviewer, back then there were only around 40 other fellow reviewers. These days, the team consists of more than 100 reviewers plus many more contributors. To have some more insight about the heavy activities, I use the mirror git repository and count the number of commits from certain email domains.

Note: commit quantity is a bad code metric (there are commits due to revert, test rebaseline, unrelated fix, etc), the amount of commits only reflects the activities and should not be extrapolated to something else.

Using the wonderful --before and --after options of git log, I can get the log messages for every commit in the last 1.5 years (18 months). After that, I use a simple script to extract the author of each commit, taking into account a patch which is committed by a different person or by the commit bot. The outcome is not scientifically correct but it may have only few false positives. Once the result is dumped (JSON formatted), a little bit magic like this:

grep author | grep -Eo '(@[a-zA-Z0-9.-]+)' | sort -n | uniq -c | sort -nr

gives me the following list (cut arbitrarily to have only 24 entries):

11712 @chromium.org
9703 @apple.com
7727 @webkit.org
2541 @google.com
1760 @igalia.com
1134 @nokia.com
1081 @gmail.com
 631 @rim.com
 378 @samsung.com
 348 @openbossa.org
 271 @profusion.mobi
 129 @company100.net
 110 @motorola.com
 100 @sencha.com
  94 @paroga.com
  91 @inf.u-szeged.hu
  88 @collabora.co.uk
  74 @adobe.com
  46 @intel.com
  42 @codeaurora.org
  40 @torchmobile.com.cn
  34 @gnome.org
  18 @stud.u-szeged.hu
  18 @forwardbias.in

Those who have access to the repository (the reviewers and committers) often use @webkit.org address instead of their affiliated organization. Some also prefer to use personal address (@gmail). Consider the importance of WebKit on mobile, do not be shocked to see more hardware-oriented companies (Motorola, Samsung, Intel, etc) are getting active. And do you spot Sencha there? :)

Now if we go back in time (good old days) and extract similar information, starting from the moment I got the reviewership early 2009, going back in time for the same span of 1.5 years, then the result looks completely different:

8062 @apple.com
3537 @webkit.org
 519 @chromium.org
 256 @trolltech.com
 216 @gmail.com
 182 @atoker.com
 137 @uwaterloo.ca
 137 @nuanti.com
 119 @gnome.org
 113 @nokia.com
  79 @collabora.co.uk
  77 @google.com
  60 @selfish.org
  53 @torchmobile.com
  43 @igalia.com
  35 @wayofthemonkey.com
  31 @gmx.de
  28 @mattlilek.com
  24 @unpluggable.com
  24 @twotoasts.de
  21 @inf.u-szeged.hu
  17 @kde.org
  16 @mac.com
  16 @belshe.com

Google Chrome was in its fancy days and thus we did not see much yet from the Googlers. Trolltech, the steward of Qt WebKit port, was being integrated to the rest of Nokia. Torch Mobile was still about to be bought by RIM. GNOME and its related consulting companies were also clearly visible.

Congratulations WebKit, and here’s to hundreds of thousands more!

  • rob

    time for a code_swarm, perhaps?

  • rob

    time for a code_swarm, perhaps?

  • Carlos

    Nice, but why the hell are Konqueror devs torturing us, users, with that crappy KHTML instead using Webkit as the Konqueror engine?

    • ariya

      AFAICS you can switch to use WebKit as the rendering engine for Konqueror. As for KHTML as the default, that’s their call.

  • Carlos

    Nice, but why the hell are Konqueror devs torturing us, users, with that crappy KHTML instead using Webkit as the Konqueror engine?

    • ariya

      AFAICS you can switch to use WebKit as the rendering engine for Konqueror. As for KHTML as the default, that’s their call.

  • Andreas

    Hey, you’re not counting the commits in KHTML before WebKit was forked, which seems a bit unfair.

    • ariya

      This is the history of WebKit only. Feel free to create an extended history covering KHTML etc.

  • Andreas

    Hey, you’re not counting the commits in KHTML before WebKit was forked, which seems a bit unfair.

    • ariya

      This is the history of WebKit only. Feel free to create an extended history covering KHTML etc.