Smart Spam protection proposal

[iShift]

In near future, when twister have 100 000+ users, twister must fight with spammers i have an idea.

User-exchanged spam list

All of us are trust our friends (followers), we can build spam protection system based on our friend opinion.

If friend see spam message it can mark it like spam, after that, this information are broadcasts to followers and how much we have spam broadcasts we can categorize spam.

we would have 3 spam category: 1) 100% spam - 80% and more of our friends mark user as spam 2) 60+% spam 60%-80% mark user as spam 3) < 50% spam 50 and less mark message as spam

Broadcasts of spam users must contain public key + username and block number when spam detected

When we receive spam broadcast we add user to our spam list and: if category is: (by default) 1) Block message and don't show any notification 2) Block message and show notification on spam page (1) 3) show message and mark it as (Maybe it is spam yes/no ?)

We can use that system for block all spam in twister: Reply/DM/Post spam

What you think about that ?

[slr]

sounds good. but what if I have 1 'friend' and what if he goes wrong with that once. will I miss that message in this case?

[iShift]

@slr if you have less than 10 friends in twister i think, smart spam may be disabled with alert "Add 9 more friends to activate smart spam"

[dryabov]

@iShift It's not so easy to befriend with unknown persons :)