I'm wondering why chapter 22 discusses locales (and now collations)
before encodings. ISTM the logical order is the reverse, because
encodings can be explained without reference to locales, but it's very
difficult to cover locales without touching on encodings. There are a
lot of forward references in sections 22.1 and 22.2 as it stands.
So I'd like to flip this around --- any objections?
regards, tom lane