It doesn't. However, the XML spec _does_ say that unless XML entities
have an XML declaration with an encoding declaration parsers are to
assume that the entity is UTF-8-encoded.
This means that if you have used ISO 8859 you may get problems, since
these characters will either be mapped to a (seemingly) random Unicode
code point or simply be invalid bit sequences that do not resolve to
any character at all.
| I guess, to correctly interpret and display those characters I have
| to know the character set which was used to encode the original text
| file.
Bingo.
| How can I communicate this character set to an XML parser?
You do this on the XML declaration, like so:
<?xml version="1.0" encoding="iso-8859-1"?>
| I would be happy if anybody could point me to somewhere I could
| start reading about this issue.
Rick Jelliffe devotes a large part of The SGML/XML Cookbook to
character sets and how they are used in XML and SGML. Other than that
I don't know of any good resources apart from good old-fashioned
digging in various places.
--Lars M.