On Sat, 1 Jan 2005 20:49:50 +0100, taliesin the storyteller
<taliesin-conlang@...> wrote:
> Demanding that the server (yep, the spec. says so) has to tell/know what
> charsets its files are in is so hare-brained that the mind she boogles
> over and rolls down the hill... (Why this is so is left as an exercise
> to the reader.)

I disagree; please explain why this is so.

For that matter, I also believe that the server should know what
*content type* its files are in and that it should announce this fact
to clients, rather than serving, say, HTML documents, PNG images, and
style sheets all as something generic such as application/octet-stream
(or, worse, text/plain). Why should charset be different?

In my website, the web server determines the content type from the
file extension on the local file system -- and it was pretty easy to
extend that to determine also the charset from the file extension by
naming things e.g. foo.html.utf8 and bar.html.l1 and telling the
server "files named .utf8 are in utf-8; files named .l1 are in
iso-8859-1".

While this is hardly relevant any longer, specifying the charset
inside the file can lead to things such as the "Netscape burp". (It
*can* be useful, though, for people who store the file locally, who
then no longer have a web server to tell them the right charset...
though this could, in theory, be added to the document on saving, just
as a file extension might be changed from, say, .php or .asp or .cgi
to .htm on saving for systems which rely on the file extension to
determine content type.)

Cheers,
--
Philip Newton <philip.newton@...>
Watch the Reply-To!