4 ms·
I imagine that the answer to that may well end up being the same as it is for email. When a comment arrives, check a few centralised reputation systems (like D
by frio 8y ago
I imagine that the answer to that may well end up being the same as it is for email. When a comment arrives, check a few centralised reputation systems (like DNSBL), and then perform a statistical analysis of the comment against some corpus (like SpamAssassin et all) before allowing it or not. We could even introduce greylisting too.
- KajMagnus 8y agoDNSBL — I'm wondering those solutions are never going to work totally well? When even Facebook gets rather many spammers, don't they. It'll be a tech race and the clever spammers can continue spamming? I'm thinking the only? way to deal with clever spammers, that's really going to work, is to manually review the first N comments by each new commenter, and only approve the comments, if they add value to the blog post. Once someone has been able to add value to the blog post 3? 5? times (e.g. provided more useful info or thoughts, or asked on-topic questions whose answers = good for others to know too) — then the likelihood that that person is a spammer is maybe 0.01%, ok to auto approve his/her comments hereafter. I actually built such blog commenting software, link in my profile.
- paulryanrogers 8y agoThat still requires a centralized reputation system to be the clearing house. Unless people frequently comment on the same blogs. And in that case the activity would be more centralized again.
- stevekemp 8y agoI setup a service to do just that: https://blogspam.net/ https://blogspam.net/