Rory Finnegan

Maybe? We currently support converting arbitrary file path types to URI types by calling `URI(p"s3://mybucket/foo/bar")`.

That seems like a great idea! The general plan I had was for folks to implement custom path types in their own packages (e.g., `WNPaths.jl`, `S3Paths.jl`, `NFSPaths.jl`) and just depend...

I would make `WindowsNetworkPath` (or `UNCPath`) a subtype of `AbstractPath` because `WindowsPath` is already a concrete type and I don't expect a lot of the `WindowsPath` methods to work for...

I'll try to reviewing it more thoroughly later in the week. Some initial thoughts are: 1. I like the idea of thinking of many filepath APIs as just AbstractTree operations....

Sorry, just forgot about this.

I think a few obvious examples include: 1. Other cloud providers like azure 2. FTP client paths https://github.com/invenia/FTPClient.jl/issues/80 3. SSH filepaths 4. Navigating and manipulating archives like zip or tar...

> The main breaking change I propose to the interface as of now is that exists should always return false on directories for key-value stores. We can provide a separate...

FWIW, I think if we were to support this we'd have the same behaviour as a collection of strings. ```julia julia> replace(("a", "b", "c"), "b" => "d") ("a", "d", "c")...

That's correct. I think random string operations should be reserved for subtypes of `AbstractString`. I also intentionally don't support `startswith`, `endswith`, regex matching, etc.

Hmm, I suppose we could special case if there's a drive, since there does seem to be precedent for that in the base Julia docs now? > If path and...