Re: OT: CXS chart and machine-readable Unicode->CXS mappings

On Mon, Mar 08, 2004 at 07:33:20PM +0100, Henrik Theiling wrote:
> Ah, ok.  I changed the charset from iso-8859-1 to utf8 throughout my
> site.  Hopefully it is better now with modern browsers.  It does not
> seem to break old browsers.
As long as you use HTML entities for all Unicode characters, the
character encoding of the page theoretically doesn't matter (I had
thought it did, but John corrected that misconception a while back).
However, in the case of Netscape 4, it *does* matter, even though it
shouldn't (possibly the source of my misconception was a dim memory of
this fact).  NS4 will not render any characters outside the Latin-1
range unless the character set encoding of the page is set to utf-8.

For maximum compatibility you should use decimal entities. This is
unfortunate, as Unicode is generally arranged and indexed in
hexadecimal; it's much easier to recognize ā than ā
as U+0101 LATIN SMALL LETTER A WITH MACRON, but the former isn't
supported by all browsers.

-Mark

Reply