HTML entities? We don't need no stinkin' HTML entities.

character_reference.rb

# coding: utf-8
#
# Encode any codepoint outside the ASCII printable range to an HTML character
# reference (http://bit.ly/KNupLT).
def encode(string)
  string.each_codepoint.inject("") do |buffer, cp|
    cp = "&#x#{cp.to_s(16)};" unless cp >= 0x20 && cp <= 0x7E
    buffer << cp
  end
end

puts encode "Japan"
# => "Japan"

puts encode "日本"
# => "&#x65e5;&#x672c;"

puts encode "Japón"
# => "Jap&#xf3;n"