Not sure what’s supposed to be inaccurate; npm allowed one way for users to break depedencies, but Elm currently allows the same way and more. It doesn’t matter how it’s done; a dependency author should not be able to simply break even projects pinned to a specific version. Go uses the same approach as Elm and recently had a prominent example of a user deleting their account, causing widespread breakage, and the username being taken over by a different user, raising obvious security concerns.
Naming policy can’t move the needle on typosquatting.
This is an unsupported assertion; making it easier to have typos makes it easier to typosquat.
elm-graphql seems like a descriptive name for a package that lets you use GrahpQL in Elm!
Not less so than a name like Graphqelm, and less so than a name that actually would describe the different scopes covered by “using”, such as code generation vs. query building vs. utility functions like dumping the schema.
Yes, and the OP in this thread is Evan’s explanation for why we shouldn’t do that in Elm.
The point was that there are other explanations to non-descriptive naming than descriptive names being taken; the latter arguably isn’t even that likely, since perfect or obvious names are a contrived idea.
This is a thread about naming.
If you’d prefer to discuss other things, I’d suggest starting a new thread!
It’s a thread about a naming policy predicated on the current flawed package management system, so the flaws are directly pertinent.
