Name: bb33257 Date: 12/14/98


Javadoc currently does not include a "charset" tag in the pages
it generates. Apparently it was thought that doing so would
interfere with the process of translating the JDK docs into
Japanese and other languages.

This is wrong. If Web browsers don't see the "charset" tag, they
typically assume either ISO 8859-1 or the current default code
page, e.g. SJIS on a Japanese system. Whenever you're generating
a web page that contains non-ASCII characters and that you expect
to be viewed by people from multiple countries, it should specify
the charset so that browsers will display it properly even if
their default is set to something else.

If pages containing "charset" tags were translated into Japanese
(say into SJIS) and the tags were left alone (so they still said
iso-8859-1), there would indeed be a mismatch. The translators
will have to change the "charset" tag to whatever character set
they used. But they should be doing this anyway. If they're
not, they're essentially assuming that anyone who wants to see
the pages has their browser's default codepage set to SJIS (or
whatever), which is not always going to be the case. For
example, I occasionally look at Japanese pages even though my
browser's default is 8859-1. Additionally, some Japanese users
have their default set to 8859-1 so that they can view the huge
number of pages out on the Web that just assume 8859-1.
======================================================================