[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: well-formedness error


On Sat, 19 Jun 2004 22:48:40 +0900, MURATA Makoto (FAMILY Given) <EB2M-MRT@xxxxxxxxxxxxxxx> wrote:

In HTTP 1.1 (RFC 2068 which is now superseded by RFC 2616), the default
of the charset parameter (of text/*) is ISO-8859-1 rather than US-ASCII.
I heard that the IESG approved this default only because that was the
real default at that time.

Nice to know. Why are we working on the assumption that 'text/*' is sent as US-ASCII, then? Isn't this part of HTTP 1.1 implemented (at all)?

Of course, this is not a real default in Asia now.

Of course not. How does things get handled over there?

When text/xml was created by RFC 2376 (a predecessor of RFC 3023) , it
was oblidged to follow RFC 2046 and RFC 2068.  That is, we were oblidged
to ensure that exisiting MIME implementations (which do not know
text/xml but rely on fallback to text/plain) correctly detect the
charset parameter.

That's OK for old clients that have absolutely no support for XML.

Thus, the use of the BOM or encoding declarations were (and still are)
unacceptable.

But why, oh, why? This won't break backwards-compatibility at all, as far as I can understand. First, assuming US-ASCII and then changing one's mind by the 'encoding' value in the XML PI would be 100% backwards-compatible, but also better future-compatible. Assuming US-ASCII no matter what the format is in, is imho a bit crazy. Why's that? Why is it that the 'encoding' value can't override US-ASCII?

Before RFC 2376 was reviewed by the IESG, the default of text/xml was
ISO-8859-1 for HTTP and US-ASCII for others.  But the IESG asked the
authors of RFC 2376 to change the default to US-ASCII.  Since there were
few implementations at that time, the IESG did not want to use
ISO-8859-1 as a default.

ISO-8859-1 doesn't help anyone, anyway (seeing that most people live in Asia, particularly). I want UTF-8 somehow. Help! :-)

--
Asbjørn Ulsberg         -=|=-        asbjornu@xxxxxxxxxxx
«He's a loathsome offensive brute, yet I can't look away»