Expand plain name fragments to unicode / IRIs

[handrews]

Currently, our plain name fragments are restricted to the US-ASCII subset of the relevant XML plain name fragment syntax. Now that we support IRIs should we expand this as well? I believe the actual XML syntax we reference covers a wider set of unicode. And perhaps we wouldn't need to spell out the syntax in our spec if we use the XML syntax exactly.

[gregsdennis]

Checking the current document, we're using IRI now. The only mentions of URIs are in the change log.

Closing.