Last modified: 2008-09-30 15:58:15 UTC
In the new Special:Statistics page, there is an unexplained "active user" count. German locaization claims "active during the las 30 days", other do not. Even if the 30 days boundary was correct (it is not) it appeary arbitrary, and I doubt that, a user who comes regularly ever 32nd day, should be seen as inactive. Imho this count is useless, wrong, and thus misguiding. It does not count currently logged-in users (We tried several times: it would not change, when we log in / out) As you can find from RecentChanges: Obviously, it does not count the number of editors, or contributors of the past 30 days. Even not if new users, bots, sysops, and anons were excluded, it it too small. Since it is hard to guess anon user identites, it is both problematic to include them in, and to exclude them from, this count. So whatever it does count, the label "active users" is incorrect and misleading for this figure. At least, something like "approximate" or "rough estimate" should be included in the wording. Something more precise would of course be better, if the figure is to be kept.
For clarification, the count is all unique editors who have performed some action (other than creating an account) however far RecentChanges goes (on WMF sites: 30 days). Bot edits are excluded. Whether this is a good metric or not is up for debate, but that's the current way of figuring it out.
(In reply to comment #1) > For clarification, the count is all unique editors who have performed some > action (other than creating an account) however far RecentChanges goes (on WMF > sites: 30 days). Bot edits are excluded. > > Whether this is a good metric or not is up for debate, but that's the current > way of figuring it out. > True. Also, the count is not generated live, and there was a bug causing it to be lower than it should be that will be fixed next sync/refresh.
Bug 41137 that is.
(In reply to comment #3) > Bug 41137 that is. > *sigh* bug 15682
see bug # 15746, bug # 15779
