Abusive sockpuppetry is an ongoing problem on the English Wikipedia and I suspect other projects as well. It corrodes community trust, and when it succeeds shifts decisions away from the best outcome. Currently we have very minimal defence against sockpuppetry, we only keep data for a very short period of time and we only investigate when we have grounds for suspicion. It would be technically feasible to be much more rigorous, and if we did so there would be big benefits in community cohesion, quality of decison making and speed of dealing with returning banned users.
- Store IP data on edits by logged in users for longer - years rather than months.
- Store data on banned user for longer still.
- Use software to automatically screen voters, functionaries and candidates for abusive sockpuppetry.
- Use Software to monitor for the return of banned users.
Reports of such possible abouse would of course need to be checked by humans first and any anomalies discussed with users by email before anything was said on wiki.
Abusive Sockpuppetry corrodes trust on the wiki and enables vote stacking in AFDs and elsewhere, if we deal with it more robustly our projects will be more harmonious trusting places with the awkward squad weakened.
Can we amend our privacy policies both to keep data on the IPs used by logged in editors for much longer, and to go on fishing trips?
Minor development costs, minor hardware costs - a few gigabytes of extra transaction data would need to be kept and a server to do some offline analysis - there would be no need to slow down any live transactions.
Do you have a thought about this proposal? A suggestion? Discuss this proposal by going to Proposal talk:Beef up our systems for detecting sockpuppets.
Want to work on this proposal?
- .. Sign your name here!
- Vibhijain 06:07, 14 May 2011 (UTC)