[discuss] Allow names to be unicode rather than ascii

[rufuspollock]

In v1.0 we restricted names to be ascii lower case primarily to support reuse in urls without escaping.

However, this is a big restriction for users with non-english alphabet who may want to name their resources and packages using their own alphabets.

Questions

• What is status of IRI type stuff in browsers and elsewhere?
• How much of a problem is this for users right now?
• What serialization of unicode do we support (utf8 probably ...)?

[augusto-herrmann]

Good question, @rufuspollock.

I think IRI have been well supported in browsers for years. According to Wikipedia, UTF-8 encoding should be used for IRIs.

I'm all for it.