Last modified: 2014-04-21 18:54:33 UTC
Although the definitions for “Active editor” [1] and “Very active editor” [2] call for counting only edits to mainspace “countable” [3] pages, geowiki counts all mainspace edits. We should make geowiki count only edits to mainspace “countable” pages. [1] https://www.mediawiki.org/wiki/Analytics/Metric_definitions#Active_editor [2] https://www.mediawiki.org/wiki/Analytics/Metric_definitions#Very_active_editor [3] https://www.mediawiki.org/wiki/Analytics/Metric_definitions#Countable_pages
Prioritization and scheduling of this bug is tracked on Mingle card https://mingle.corp.wikimedia.org/projects/analytics/cards/1205
I am not so sure about this -- countable namespaces was introduced by Erik Zachte without a public discussion (AFAICT) and I think that myself and Dario have different thoughts about this as it makes (among other things) cross-wiki comparisons harder. Adding Dario to ticket.
(In reply to comment #2) > I am not so sure about this -- countable namespaces [...] ^^^^^^^^^^ I was referring to countable /pages/. Do we have the concept of “countable” also for namespaces? Where is this defined?
We do have a concept of a "content namespace" that varies across wikis. http://www.mediawiki.org/wiki/Analytics/Metric_definitions#Content_namespaces My opinion is that both "countable pages" and "content namespaces" are MediaWiki artifacts that we cannot and shouldn't include in our KPI definitions going forward. 1) They make historical analysis pointless (as individual communities can change these definitions over time) 2) They make cross-project comparisons pointless (different projects use different definitions of "countable" and "content") 3) They make feature-level -> project-level projections impossible I recommend that: • we start using a stricter definition of active limited to ns0 and including all pages, not just countable ones. • we discuss on an ad-hoc basis how to accommodate special projects like Commons or Wikidata • we re-evaluate the pros and cons of this approach after a few months
(In reply to comment #4) > I recommend that: > • we start using a stricter definition of active [...] From my understanding, geowiki is a production service. I do not like the idea of experimenting with new definitions on production services. Instead I prefer to make it agree with our team's commonly accepted accurate definition. Otherwise, we can never compare numbers between services. So which definition of "active editor" is now accurate? The most recent statement that I have is from our product manager from Sept 20, where he stated that [1] actually is accurate for "active editors". Coincidentally, that was even in geowiki context. So if that hasn't silently changed in the meantime, let's use this bug to track our progress to get [1] implemented in geowiki. [1] https://www.mediawiki.org/w/index.php?title=Analytics/Metric_definitions&oldid=689732#Active_editor
Christian, you have a good point that this is not the appropriate context to make that decision. For the record, the definition you refer to is neither prescriptive nor "the team's commonly accepted accurate definition". It's a definition that describes the way in which we generate the TAE metric for wikistats/the monthly reportcard. To my knowledge, this definition has never been used by any team in Product or Editor Engagement, i.e. those teams that have been dealing for years with active editor numbers as part of their mandate. I suspect the same applies to most other programs working on editor engagement initiatives in the org. Let's move forward with [1] in geowiki, I'll start a separate thread to discuss the issues with the definition of active editor.