XWiki Old Project Health

Last modified by Vincent Massol on 2024/10/30 11:09

Downloads

Downloads for the WAR + HSQLDB packagings:

XWiki Downloads (WAR+HSQLDB)

Source:

Interpretation:

  • At one point we had more distributions (workspaces, xeclipse, xoffice, xwatch, etc) and this could have lead to more downloads
  • Since XWiki 3.3 we now have a Debian distribution (.DEB), installable through APT and we don't count those downloads in the figures above. They could account for an important part, we would need to retrieve their stats.
  • Since XWiki 8.4 we now have Docker images too.
  • These download figures need to be correlated with the Active Installs.
  • These downloads do not count XWiki Extensions.
  • The majority of XARs are downloaded from the Extension Manager and are not counted in these stats.
  • In 2018 we stopped publishing our binaries on OW2 and thus we don't have the download figures since then. In addition, since we have several distribution channels (APT, Docker, xwiki.org, etc), it's impossible to track the download numbers. Thus we're now relying solely on Active Installs figures which provides a better tracking metric.

Corrected Graph:

We're using the Active Install Distribution proportions to correct the download figure above. 

Note that the Distribution data are only retrieved for XWiki 8.4.4+ so the corrections make sense only from 2017+ and they only been applied from 2017 forward.

XWiki Downloads (Total)

Interpretation:

  • Based on the corrected graph, we're actually increasing the number of downloads of XWiki overall!

Mailing Lists Activity

Source:

Interpretation:

  • Either less users since July 2010 or the product is easier to use
  • Either less devs since July 2010 or they're asking less questions (more proficient?)

Mailing list Traffic

Commit Activity

Commits for Platform, Enterprise, Manager, Commons, Rendering, XOffice, XEclipse, XWatch, XWorkspaces:

Commits

Commits for Contrib:

Commits

Source:

  • SVNSearch and Git/GitHub
  • To compute all commits between 2 dates in Git: git rev-list HEAD --count --since="2011-06-30" --until="2012-06-30"

Interpretation:

  • See also the Commit Activity on Ohloh which also shows a small commit slowdown since 2011 (i.e. since we moved to Git)
  • Starting in 2011 we moved to Git (we were using SVN before) and this could have caused some difference in # of commits. For example sandbox and contrib code was moved in different Git repositories. In addition it's possible that using Git generates less commits because with SVN merges were a pain and thus devs were committing very frequently to not have to perform merges... emoticon_wink
  • Some projects have been deprecated and moved to contrib such as XEclipse, XWatch, XOffice, Workspaces, XEM, etc and this has probably resulted in reduced commits
  • We had 7605 commits in 2011 means about 20 commits/day. This is to be compared to other well known projects in 2011:
    ProjectCommits per yearCommits per day
    Gnome36500100
    Linux47450130
    Drupal48545133
    KDE109500300

    Classification:

    Category
    XtraLarge
    1000+ devs
    100+ commits/day
    Linux, KDE, Apache, Eclipse, Perl+CPAN, Drupal, Gnome
    Large
    20-200 devs
    50-100 commits/day
    MySQL, Qt, OpenOffice, Mono, JBoss
    MediumGIMP, SVN, wordpress
  • Comparison between XWiki, Foswiki and MediaWiki on Ohloh

Committer Statistics

Source Contributors

committers.png

Source:

Analysis:

  • For the full list of committers and their stats see also the Hall of Fame.
  • Seems the number of contributors has been steadily increasing till mid 2011 and from then on it's pretty stable between 8 to 15 different contributors each month.
  • Thomas Mortagne is increasing his lead and is moving towards a 30% commit share for all times (27% ATM and 30% for 2015)
  • Sergiu has reduced his participation, that's too bad! Come back Sergiu emoticon_smile
  • Jean-Vincent has left the project and is no longer active unfortunately
  • Vincent is reducing slightly from 20% all time to 17% for 2015
  • Marius is stable at 9%
  • Guillaume is increasing: from 2% all time to 15% for 2015
  • Eduard is also increasing: from 2% all time to 10% for 2015
  • Pretty stable in # of core committers (but this doesn't take into account the activity on contrib! We need to find a way to analyze this).

Source Repository Statistics

  • Sonar provides lots of stats of the XWiki projects over time (for modules with Java code). Note that the Test Coverage Percentage doesn't take into account integration and functional tests (thus if a unit test also tests code from another module these stats are not taken into account).
  • SVNSearch provides some very nice stats
  • Fisheye also provides some stats
  • Using Ohloh

l10n Activity

l10n modified translation keys

l10n Contributors

Source:

Analysis:

  • User "114 Hareketi" has 21K in from 2013-06-30 to 2014-06-30 but seems he saved 6 times several keys he modified... We would need to have unique modifications per key per user.

Releases

XWiki Enterprise/Standard Releases per year

Source:

Analysis:

  • At some points we were releasing other projects in addition to XWiki Enterprise (XWiki Office, XWiki Enterprise Manager, XWiki Watch and XWiki Eclipse).
  • Pretty stable release cycle which means we're globally following strictly our release plans at a rate of about 30 releases per year, that's 1 release every 12 days, or 2.5 releases per month. Our release strategy is to release a minor version every 2 to 2.5 months in average. However we do release milestones, RCs and bugfix releases too.

Analysis

2010

XWiki in 2010

2009

XWiki in 2009

General Stats from SourceKibitzer (1/1/2004 and 1/11/2007)

This data comes from the old SourceKibitzer (now Programeter) product and were obtained through XWiki's SVN. The graphs were generated from the data below (1/1/2004 to 1/11/2007).

xwiki-statistics.png

Legend:

  • Contributors - different repository commiters
  • Size - Size of the codebase measured in NCSS (non-commenting source statements)
  • Activity - Number of changes during the last 3 months. By change here we don't mean commit, but the change in a single file.
  • Complexity - Average complexity of one non-comment statement
  • Density of comments
  • Stability - percentage of codebase being stable during the last 3 months

Analysis

  • Committers are increasing steadily
  • Mailing list traffic is increasing fast
  • Activity is increasing
  • Complexity isn't increasing
  • We don't have enough comments in the source code (should be at least 50%). However for the past 3-4 months we have been steadily increasing the comments so we're progressing in the right direction.

Data

DateCommittersSizeActivityComplexityDensity of commentsStability
01/01/04123231120.290.010.07
01/02/04163873900.290.020.11
01/03/04180566820.30.020.11
01/04/041124829350.280.020.25
01/05/041145919310.290.030.4
01/06/041161917970.30.030.45
01/07/042173616730.30.030.62
01/08/042178644590.30.030.66
01/09/042192853850.30.030.65
01/10/042197092370.30.030.76
01/11/042199481950.30.030.78
01/12/042211091870.310.030.83
01/01/052219731370.310.030.82
01/02/052224641650.30.030.82
01/03/05322503920.30.030.89
01/04/05423051820.30.030.89
01/05/0572471012370.30.030.51
01/06/0591166216960.220.060.75
01/07/05101316318180.320.050.75
01/08/0512152156500.30.050.74
01/09/05174428210240.290.050.15
01/10/05204991215420.30.050.15
01/11/05205021415600.30.050.16
01/12/0521503157360.30.050.75
01/01/062150335660.30.050.97
01/02/062250374230.30.050.99
01/03/0625516955210.30.050.92
01/04/0625533197190.30.050.86
01/05/0625537167810.30.050.85
01/06/0625540093010.30.050.92
01/07/0625541051100.30.050.96
01/08/0625950136100.30.040.6
01/09/06269874613960.30.040.54
01/10/062610085418680.30.050.52
01/11/062910507125730.30.050.67
01/12/062910583221380.30.050.68
01/01/07296771022640.310.070.69
01/02/07297097012410.310.070.87
01/03/07317186313650.30.070.91
01/04/07317553812590.30.070.87
01/05/07317673014440.310.070.84
01/06/07327956618180.310.070.84
01/07/07328449443390.30.080.52
01/08/07328654645610.30.080.52
01/09/07349214859690.30.090.5
01/10/07359696343370.310.10.61
01/11/07399860438850.310.10.62

Get Connected