Re: ALT revisited (short-term)

to follow up on what Daniel Dardailler said:
> 
> > What I was suggesting relies on the fact that the header fields
> > exchanged in HTTP can carry metadata about the message body.  A
> > server can provide a description of the image as the value of a
> > Content-Description: header or indirectly via something like an
> > X-Content-Description-References: header.
> 
> There is no Content-Description in HTTP1.1 so this would have to be an
> extension.
> 

I have not re-read the 1.1 document on this point.  What it should 
say is approximately:

	"The HTTP messages shall [follow the 822 format and]
	contain the following header fields required for
	managment of the communication process:
		<list of clauses>
	The HTTP messages should also contain all header fields
	required for the Internet Media Type [cite STD] identified by
	the Content-Type: header.  The HTTP messages may include other
	header fields conforming to the extension practices defined in
	Internet Standard 7?, RFC 822."

The HTTP document should not interfere in the separate management
of media types, which has an established, orderly process for
evolutionary upgrade.

What I'm trying to say is that syncronizing the usage of headers
appropriate to different media types should not require a change
to the HTTP transport specification.

It does require action within the types management discipline, but
there is an established process there for us to ratchet our 
recommended practices from X-experimental to registered
to standard headers.

--
Al Gilman

Received on Wednesday, 2 July 1997 10:49:36 UTC