Hi folks,
As we've gotten down to brass tacks in planning the deployment of Flagged
Revisions on
en.wikipedia.org, it's become obvious that we'll probably need
to limit the number of articles put under Flagged Protection at first. A
couple of reasons for doing this:
a) Performance - we know already from
de.wikipedia.org and from watching
flaggedrevs.labs.wikimedia.org that this feature can impact performance.
So, we'd like to start small and build up from there
b) Community norm development - we'd like to give the community a chance to
use this in production in a limited way to start with to get a feel for the
feature in the wild. We suspect many people will appreciate the opportunity
to see this in action and get a better sense of the policy implications
without having to worry about finding that half of English Wikipedia is
under Flagged Protection.
So, we're planning on putting an upper bound of 2000 articles when we start
the trial. We'll then see how this plays out. If performance takes a
severe hit, we'll need to work with the admin community for a plan to back
down from that number (in lieu of total reverting the feature). If things
are going well on all fronts, we can possibly bump things up.
I've outlined this here:
http://en.wikipedia.org/wiki/Wikipedia:Flagged_protection_and_patrolled_rev…
What would be useful for this group to figure out is:
1. What articles would be the best candidates to start off with? From a
technical perspective, it'd be really handy to have a representative sample
of traffic characteristics (e.g. high traffic and low traffic articles),
since caching is one of the big areas of performance hit.
2. If we need to start weeding out articles to get the total count down,
what order should we go in?
Please leave your thoughts on the talk page, and/or flesh out the portions
of the main article that still need love (that page in particular needs some
help).
Thanks!
Rob