> So what’s a bad actor? I consider someone who squats a bunch of crates to make or point or prevent their names from being used to be a bad actor. In my exper

> So what’s a bad actor? I consider someone who squats a bunch of crates to make or point or prevent their names from being used to be a bad actor. In my experience, the thorny problem with namespaces isn't bad actors because you can fix that with content moderation. Yes, that takes time and effort which can be scarce in an open-source community, but it's doable. No, the thorny problem is good actors with bad execution: people who publish a package with a nice name that does what it says it does, more or less, but does it poorly. It doesn't cover all the edges case or it is poorly optimized or it has limited features or it has some fatal design flaw in its architecture... Maybe it was fine at some point but just hasn't kept up with the times. So what do you do if a package like that grabs the name "json" or "http" or something? You don't want to kick packages off just for not being great, but you also don't want it to seem like the de facto package for a task in your language. I do a lot of Haskell, and it's not the best developer experience when you have to tell people "don't use json, use aeson instead". Namespaces would solve this problem. You could have a "blessed" namespace (like rust/) and, when functionality becomes important enough that the quality of the "default" package matters use the namespace to provide that package (rust/json, rust/http... etc). You don't have to kick off the existing package (even if it's squatting a "nice" namespace like json/json) but you also don't get into the point where packages with nice names look fine but are actually flawed and a bad choice for real work.