MZ brings up a good point. Do we have any idea what the reject rate is
for Abusefilter on desktop?
Also is there any way to view the edits that triggered AbuseFilter to
get an idea bout what % of them were actually vandalism?
On Mon, Aug 19, 2013 at 6:25 PM, MZMcBride <z(a)mzmcbride.com> wrote:
Jon Robson wrote:
Between 31st July 2013 (00:00:00) and August 19th
(@ 18:53:18)...
15089 of all mobile edits resulted in an error compared to 23455
successful edits (38544 edits in total)- that's a 39% error rate which
is simply unacceptable. The breakdown of these errors is as follows.
The most alarming is AbuseFilter - it is accounting for 72% of all
editing errors, costing all of our projects a lot of edits.
[stats]
These numbers don't mean a whole lot (to me, at least) without a
comparison to stats for non-mobile edits. We knew when the AbuseFilter was
introduced that it wasn't always going to be used to filter only abuse.
People naturally use it for all kinds of crazy purposes these days.
Without a bit more context, it's difficult to know what's actually a
problem (in relative terms).
MZMcBride
_______________________________________________
Wikitech-l mailing list
Wikitech-l(a)lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l
--
Jon Robson
http://jonrobson.me.uk
@rakugojon