That's a good point, the block stops them from replying altogether (or quote tweeting).
I think the fundamental property of a block feature should be complete undetectability that the user has been blocked by the user in question. For example Usenet killfile: you never know if you're in somebody's killfile unless they tell you so. Twitter's block function obviously fails this test, which makes it completely useless: once you realize you're blocked, just create another account to continue harassing the user who blocked you.
Mute does that.
So you're saying Twitter already has a block function that actually works? Then it's all the more baffling they also have one that doesn't work, and yet is somehow more well known.
They have 2 different functions with different names that do different things. They call the method of blocking you prefer "mute".
My point is that "mute" is strictly better than "block", so there's no reason to have "block" in the first place. Even if somebody likes the misfeature of the blockee knowing they're blocked, you can always tweet "I just muted @theannoyingperson", so that @theannoyingperson knows they're muted.