Ah, but to hell with rationality! You think it's rational for someone to reply to all of your tweets with mindless abuse, and to encourage other people to do the same?
For better or worse (I think it's for worse, which is why I don't tweet all that much), Twitter encourages rapid-fire think-before-speaking responses to things. You see something you don't like, you hit reply and boom, you've just told that person to die in a fire. Or you've just retweeted it to your small but equally annoying group of friends. The purpose of the old blocking system, in a way, was to prevent trolls[1] from seeing things that might provoke them into action.
The goal is not actually to prevent someone from seeing your messages, but to prevent someone who is logged in to an active twitter account with a >0 number of followers from seeing your tweet. That's a non-trivial difference. I started out (in my other comments) thinking that the difference really was trivial, but you actually have to think like a troll in order to understand why it isn't. Unless they really cared about harassing someone, the old-style block would probably be enough to stop most people from tweeting an insult, if only because it forces them to 'cool off' by logging out, finding the person they want to harass, and logging back in again in order to reply.
I can't shake the feeling that neither policy is really right, and there must be a more elegant solution. But, it's 1am here and I should probably get some sleep before trying to find the solution that has evaded the entire internet for 20 years.
[1] I don't use the term 'trolls' lightly. Lots of things get called 'trolling' which really aren't, but in this case I think it's valid.