My view is that Twitter could solve many of the problems by creating 2 scores out of say, 1000, for each user. Given the millions on Twitter a score out of 1

My view is that Twitter could solve many of the problems by creating 2 scores out of say, 1000, for each user. Given the millions on Twitter a score out of 100 would be too coarse. 1. Don't censor anything from anyone who I follow. 2. Score A is a metric based on geographic location similarity to you, and who you follow. So if you are bicoastal and follow your friends from LA, NYC you won't have Oregon or Pennsylvania proximity bumping up someone's score. But a person in Ohio whose friends cover the tri state area will have a different score for the same twitter account. 3. Score B is topic similarity and popularity. So if an account is sharing news about e.g. Docker/Mesos/k8s, and it gets a lot of likes or retweets, then it has a higher score. Finally, you put an option in the user settings to block someone commenting on your TL if score is less than x , and silently down rate if score is less than that. For both scores A and B. The above is based on what I have seen on Twitter, which is unknown to you people doing a sort of drive by harassment, often after someone who disagreed with you retweets your post.