The XML decl is not a PI (especially in SAX)

Rick Jelliffe ricko at
Sun Nov 14 16:39:20 GMT 1999

The XML header could be a PI tag based on its delimiters, but it is not
by definition. James Clark has picked me up on this a few times: he is
so tough. The XML header is stripped as a function of entity management.

However, even though the encoding attribute in the XML header does not
form part of the information set of  a document, I think it is a mistake
to think that it should be discarded after the parsing of a document.

On the contrary, I think applications should operate so that, by
default, the output encoding of documents matches the input encoding (of
the root entity, unless the root entity is ASCII and subentities use
some other ASCII-superset encoding, in which case that encoding should
be used.)

I hope that future APIs could keep this in mind: hide the encoding from
the processor by all means, but save the encoding information for
re-encoding later.

Rick Jelliffe

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at
Archived as: and on CD-ROM/ISBN 981-02-3594-1
To unsubscribe, mailto:majordomo at the following message;
unsubscribe xml-dev
To subscribe to the digests, mailto:majordomo at the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at

More information about the Xml-dev mailing list