Re: Proposal for an HTTP ERR method

Jamie Lokier wrote:
> I don't have a position either way.  I suggest that if RFC 3023 should
> be obsoleted, it is should be only if there's an abundance of clients
> which look at the <?xml...?> declaration given "text/xml" -- in
> effect, giving up a requirement of RFC 3032, in the same way that HTML
> 4.01 says to give up a requirement from RFC 2616.
> 
> I don't know if there is an abundance of such clients.

Btw, if I were writing a new client, prior to this thread I would have
made it interpret "text/xml" as meaning to look in the XML declaration
for the charset.

Not out of maliciousness, but ignorance: it had simply never occurred
to me that "text/xml" meant the same as "text/xml; charset=us-ascii".
I've read the XML 1.0 documentation quite closely, and I still managed
to skip over that bit.  I.e. I didn't read RFC 2376 properly which it
references, because it didn't occur to me it was important in this regard.

It's also a bit counterintuitive that an XML file which specifies a
charset in its XML declaration would have one meaning when received as
"text/xml", and another when saved as file.xml and then read again.
Even on a filesystem which saves the content-type as a special
attribute, that's the case, as XML 1.0 says:

   In the interests of interoperability, however, the following rule is
   recommended.  If an XML entity is in a file, the Byte-Order Mark and
   encoding declaration are used (if present) to determine the character
   encoding.

-- Jamie