I *like* having a language tag associated with a UTF-8, and this gives
implementors the same option as before but allows them to include a language
tag.
The current direction we're supposed to be pursuing is language tags embedded
in UTF-8. The facilities for this have received approval by the Unicode
Technical committee and now await approval by the ISO 10646 folks. And there is
now an Internet draft by Glenn Adams explaining how this works.
Given the current state of affairs there should be no need for a language
tag associated with UTF-8-only strings. Strings which allow for different
charsets are another matter, of course.
Ned