(Untitled)

[onah]

I just got some spam that contains text which makes me strongly suspect that some person or spider is pulling text from my LJ and sending it to me in a spam (along with girl pictures or something that mutt doesn't display), in order to get past my spam filters

Comments

[jope]

Addresses known to harbor spammers is less and less a meaningful statement, as the scavenging of this text and sending of spam is increasingly done by hijacked PCs. ISPs should be doing egress spam filtering to counter this, but there's no profit motive for them (or at least not one that I've seen widely accepted), so generally they do not.

[onah]

so, LJ should block http requests from ISPs that get in the RBLs. shouldn't that be possible? that would provide a profit motive.

[jope]

You're missing the point: No ISP can prevent 100% of its users from getting hacked and used as zombies. And since most use is through dynamic public IPs, there's no easy way to backtrack to any individual user, based on just the IPs that got them landed on an RBL. IP plus timestamp of the infraction would work, but I doubt that most RBLs require that level of granularity, unfortunately.

One approach that holds some promise is penalizing ISPs that do a worse job policing their users, not by denying them access to the site altogether, but jacking up their latency, reducing bandwidth, etc. In practice though, that can end up biting the site since that kind of throttling drags out usage of certain resources (e.g. network connections, database connections, server thread/process) even though less is being done with them.

[goldfish42]

The best you can do with LJ is turn on the "block spiders/bots" option, and it will catch some of them.