Allow community maintainers to subscribe to other communities' ban lists.

[netp_npokon]

Title
Allow community maintainers to subscribe to other communities' ban lists.

Short, concise description of the idea
Give the community maintainers the ability to share their ban lists with other communities' moderators.

Full description of the ideaLivejournal suffers from hordes of bots that post scam into multiple communities. The text and links

Comments

[lady_angelina]

I'll have to say no on this one, due to the privacy considerations. Yes, the ban lists may contain a number of bots... but community maintainers also ban users for reasons having nothing to do with trolling or spamming, and even if a user caused trouble in one comm, that doesn't mean they should automatically be blacklisted at another.

[azurelunatic]

I would be totally down with this except for a few things.

1) Spam isn't the only reason for banning -- the user may have violated a rule unique to that community, or angered a maintainer. To fix that, there would have to be a "ban for spamming"/"ban for some other reason" option for banning.

2) The suggestion does not say what the source community would have to do in order to share their ban list. Would it be a choice of a source community maintainer, or would anyone be able to subscribe without the maintainer changing a setting? If it is a choice, I am all for it. If anyone would be able to subscribe to the ban list of any community without the maintainer saying something, I would have to say no.

3) What about un-bans?

[azurelunatic]

4) If this were implemented, and the suggestion for communities being able to set notes were implemented as well, it would be good to automatically set a note on the banned user as "banned from [community] spam ban list" and the date or some such thing. (This would be only for paid communities

[lady_angelina]

#6 is what I meant by in my comment, yeah. Thanks for explaining it worlds better than I could have.

[charliemc]

Agreed.

On some levels I can see how helpful this would be, but overall I'm nervous about this suggestion...

[lied_ohne_worte]

6) This could potentially lead to chains of banning: community A bans a user, and community B and C who are subscribed to community A ban them too. Community D is subscribed to Community B's ban list, and therefore Community D bans them. Community E is subscribed to Community D, and bans them.

Exactly. While I think the idea isn't necessary bad, it could lead to lots of confusion. More user than you'd think have been banned from comms for a variety of reasons, with the ban reason often being something that another comm does not care about, and cross-subscriptions could lead to users being banned without knowing for something that the comm they are banned from does not consider ban-worthy.

[elyssa]

I think it's an idea that has merit, but that the implementation you describe wouldn't necessarily work out so well.

What I could see as working if you want to specifically target spamming is that if a user has had the exact same post/comment deleted and marked as spam by, say, more than some arbitrary number of different community maintainers then... well, something.

There's already a spam reporting system, but it's currently only partly automated. It would be interesting to open discussion on actually banning someone with more than X number of identical posts/entries that have been deleted and marked as spam from posting in communities for a certain amount of time. Or some other deterrent from making that type of posts.

Of course, then there's the problem of people who'll just make a large number of accounts and use those. Still, there's some food for thought here.

[azurelunatic]

I could see the current "cannot comment or create an entry" state that unvalidated users are in being used as not quite a deterrent, but a damage control thing. Perhaps have a certain threshold number of reports to activate it, then have time added on to the damage control timeout, and then find a sweet spot for that time so a user that's likely to be a genuine spammer would remain unable to cause further damage until suspended, and someone who was not actually a spammer would merely get a wakeup call.

I should think that closing a spam report (or all spam reports) would dismiss the user's spamwatch status, as that would establish that someone had their eye on it and they were not in fact a spammer. (Well, closing the spam reports and suspending them would also take care of things.)

[mskala]

Any spam-detection system that relies on detecting the "exact same post/comment" will be defeated by the random gibberish spammers already add to their output.

[azurelunatic]

This is true, and spammers would learn that quickly. However, I've encountered a surprising number of identical spam comments while wearing my Dreamwidth volunteer hat, and a leaky bucket is better than none.

A better option would be a smart, trainable system.

...huh. Hook that tool that Tupshin was using to clean out latest.bml up to the spam queue and see what it does.

[mlady_rebecca]

Interesting idea, but I don't think I'd want it fully automated, even if there was a way to separate bot bans from bad behavior bans. I'd rather see the list of bans and have the option to ban each user one by one.