I've been reading the HTTP and HTML specs about character sets. The HTTP spec says, "If a page is of type 'text/*', and the HTTP headers don't specify a character set, assume ISO-8859-1'.
The HTML spec says, "Don't follow the HTTP spec rules about the default being ISO-8859-1", and "Use the HTTP-specified character set first, if any, and after that the character set specified in the HTML itself with a <META> tag". Though I personally think that the in-document character set specification should override the one specified in the HTTP headers, I'm following those rules for HTML. Does anyone know of any interesting problems with those rules? Bill