Im not sure i completely understand your question, but this may help:
http://www.cl.cam.ac.uk/~mgk25/unicode.html#web
http://www.utf-8.com/

Carl.

The problem of language detection is quite difficult. Instead of forcing
a browser to use heuristics to determine document language, an HTML tag
can easily announce the same information. See the HTML "lang" tag:

http://www.w3.org/International/tutorials/tutorial-lang/

Regards,
John O'Conner

That is the purpose of the "charset=" clause in an HTTP "Content-type:"
header.  There are various ways to put such a header into your response,
but this is probably the easiest:

(in servlet code:)

    response.setContentType("text/html; charset=UTF-8");

You do that before starting to write the actual response text.  You then
ensure that you actually do encode the text in UTF-8, either by getting
and using the response's Writer, or by wrapping the response's
OutputStream in an OutputStreamWriter configured to use UTF-8.

On Fri, 24 Feb 2006 22:26:53 GMT, "Shinya Koizumi"
<xxxx@smartestdesign.com> wrote, quoted or indirectly quoted someone
who said :

It could have a BOM. The browser could set fields in the header to
give the server a hint.

See http://mindprod.com/jgloss/bom.html
http://mindprod.com/jgloss/http.html

"Shinya Koizumi" <xxxx@smartestdesign.com> wrote::

Inferring a language from the characters contained in any given document is
difficult work, even for humans. One could write some heuristics to make
guesses, but those guesses will often be wrong. You don't say how much being
wrong matters for your application.

-- Adam Maass