re: Determination of Encoding

• From: Gavin Nicol <gtn@eps.inso.com>
• Date: Tue, 17 Jun 1997 10:19:54 -0400
• To: w3c-sgml-wg@w3.org
• Message-Id: <199706171419.KAA28132@nathaniel.eps.inso.com>

Tim writes:
>It is uncontroversially good that processors should use whatever
>heuristics work in order to figure out document encodings.  It is
>controversial, but the result of a strong majority decision, that 
>XML-lang should include an in-band signaling mechanism.
>
>I have hear any convincing argument why the spec should try to clean
>up this mess single-handed, and why it should say one word more than
>it does today.

I would like to have some mention of the possibility of using
meta-data in Appendix E. Something like the following would suffice:

  I cases where it is possible to determine character encoding
  from meta-data (for example, HTTP headers), this information 
  should be used. It is a reportable error for the meta-data and XML
  declaration to differ.

Anything beyond this is unnecessary.

Received on Tuesday, 17 June 1997 10:20:34 UTC