Another problem with the flagging system

[madfilkentist]

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

Comments

[cdaae]

Regarding comment harassment, there's a process that will get the APT to do something. Basically if it happens, you have to turn off anonymous commenting, turn on IP logging, and ban the username/s doing the harassing. Then if/when they use another username to do it again, report them for circumventing a ban. Unless the APT claims they can't be sure it's the same user, they will generally give the user a "notice of no contact", meaning that user isn't allowed to leave you comments, or leave comments about you, or refer to you in any way anywhere on LJ. If they break it, they get suspended.

The flaw in the system is that someone who knows how to use proxies and to disguise who they are can get around it.

[archmage]

That sounds like it shoudl have been less a "flag this" situation than a "directly report the harassment" thing.

The flag system isn't supposed to be the be-all and end-all of systems, it's merely one tool, and, as it's still new, I expect the tool to get the kinks worked out and be refined as it sees real-world use.

[pandorasblog]

It says a lot about what flagging is really for; ie discouraging people from posting stuff that violates the groupthink that LJ considers conducive to good publicity (i.e. nothing you wouldn't want a primary school child to see).

Xanga had the very same problems with both individual harassment cases and flagging. I don't feel so at home there nowdays.

[lied_ohne_worte]

The flagging system refers to the content of entries, not to comments. It is not intended to function in such cases, which is why you can only flag entries, not comments. Instead, the offended user can open an Abuse request with the appropriate links, and it will be investigated. She will not need additional people to back up her complaint. If your friend hasn't done so yet, you may wish to point her to this link http://www.livejournal.com/abuse/report.bml - as you can see, it contains a category that appears to be what her problem is about (i.e. a comment in her journal).

The rule that entries have to be flagged by several users in order to be reviewed is necessary, I suppose - otherwise, a troll could sit down for an hour, flag several hundred posts and force the APT to investigate them all.

To sum it up - while I understand that you are concerned for your friend, your perception of the flagging system isn't entirely correct.

[mementomei]

How hard is it to make a few (or 100) sockpuppets from several different fake email accounts, then flag away from all of them? The LJ flagging system is woefully flawed. Someone with an axe to grind could shut a legitimate user right down. I believe that certain categories of flags get the account suspended automatically. (I thought I read that--not sure.)

[lied_ohne_worte]

I believe that certain categories of flags get the account suspended automatically. (I thought I read that--not sure.)

They won't suspend anyone without investigating the reports they received; also, users can be suspended for the same stuff for which they could already be suspended before. Also, users reported others to Abuse out of grudge before the flagging system was implemented.

I'm not saying that the flagging system is perfect; what I wanted to point out was that the OP had some misconceptions about how it works and what it's supposed to do. There are so many rumours and half-truths around already that I think especially in this comm, which tries to be something like a voice of reason, we should try not to spread paranoia and misinformation.

[qfemale]

Accounts need to be older than 1 month before they can flag anything. So even if somebody does bother to make 100 accounts with fake email addresses (and you still need to verify your email at some point before you can use your journal to 100%...) you'd still have to wait a month before you can flag anything.