Yet unanswered questions

  1. How extensible should the grammar be? Two stakeholders:
       reading device manufactuers: predictability/control of behavior
       the information: structure/semantics/mediatypes that best represent the content
  2. How should a document define which extensions it uses?
       Reading devices need this information (?).
  3. What behaviors should be expected from a reading device regarding extended/unknown content types?
  4. How complex can the extension mechanism be?
    To a large extent, this depends on who does the extension:
       in a specification phase: complexity a lesser problem
       in a production phase: complexity a large problem