Another problem with the flagging system
A recent entry on lj_policy shows that not only is the flagging system prone to abuse, its design makes it difficult to report legitimate problems. The post states: "We'd like to reiterate that there's no effect at all from a single user flagging another user's content. After multiple flags, the content will be reviewed."
Recently someone I know was subject to serious comment harassment on her journal. I mean really serious stuff: pictures of mutilated dead bodies, accompanied by insults. This is the kind of thing which should get a user booted, since it amounts to an implicit threat of violence. But when I tried to flag the comments, I got a message telling me I can't flag harassment of another user; she's on her own. But her own complaint, unsupported by others, will not even trigger a review! The offending user has not been terminated.
Flagging was doubtless set up as a labor-saving mechanism; it's easier to read a bunch of standard forms than a bunch of e-mails which are all differently organized and may be badly written. But when it favors the reporting of material which "offends" a pressure group while making it impossible for individuals to get attention about serious problems, that's broken.