I think you are saying that it would be OK with existing practice if:
    - the document character set is IS 10646
    - the default representation character set is ISO-8859-1
Yes.  Though I would prefer to express the second point as something like:
    - in the absence of an explicit CHARSET parameter in the HTTP
      Content-Type header, a CHARSET=ISO-8859-1 is to be assumed.
Glenn