Re Risker's concern at the new page patrol problems being replicated at flagged revisions. I agree that there are problems at new page patrol, including both excessive speediness where pages are judged and tagged within seconds of being created and occasional underkill where plausible looking hoaxes are tagged as patrolled.
The underkill problem simply means that this is not perfect and we will still need other methods of spotting vandalism such as watchlists and looking at vandals previous edits. The excessive speediness problem I would hope will be less serious with flagged revisions as it merely establishes the principle that every edit needs to be a net positive one. In my view it is a problem if an editor creates a page with a good faith line or two and then has it deleted before they can add the next line. I'm hoping that overkill will be less at flagged revisions, and more easily dealt with. One reason why I think this is that my experience has been that hugglers and others who revert vandalism tend to be quick to apologise and revert mistakes when IPs and newbies ask why their edit was reverted. But tagging of one line articles for deletion is less likely to be challenged by newbies or seen as an error by new page patrollers (I've had lots of good feedback from new page patrollers for other CSD tags I've declined - but rescued one line articles can get the response "nicely rescued, but it wasn't like that when I tagged it"). I think that flagged revision patrolling will work best if we learn the lessons of our existing systems and think of it as a series of sieves of different grades. An initial very coarse grade that identifies a lot of obviously good edits such as typo fixes and obviously bad edits but leaves the complex minority flagged. A few days for watchlisters to revert subtle vandalism and mark some good edits as good. Followed by some more detailedl checking by those who work at the back of the list on the oldest edits still flagged. One enhancement that may seem counterintuitive would be to fix the back of the list at a short interval such as a week, with anything not flagged after then going live at that point. This would mean that everyone who checks their watchlist weekly would have had a chance to revert vandalism before it went live, and would mean that the wait would never be longer than 7 days. WereSpielChequers > Given that New Page Patrol is constantly at a backlog of between 27-30 days > (that is, there are always a significant number of new pages of that age), > while at the same time we have problems with new pages being patrolled *too > quickly* and CSD'd within 2 minutes, I think we will see the same issue with > flagged revisions: that is, some edits being quickly passed without proper > review, allowing sneaky vandalism in, while others take so long to be > reviewed it takes away the wiki flavour. > > On the other hand, it might be a very different way of managing edit > warring. > > Risker _______________________________________________ WikiEN-l mailing list [email protected] To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
