Re: Refresh, non-standard headers, [was Re: Forcing to another page to load]

> Date: Mon, 7 Jul 1997 10:10:15 PDT
> From: Larry Masinter <masinter@parc.xerox.com>
> To: http-wg@cuckoo.hpl.hp.com
> Subject: Refresh, non-standard headers, [was Re: Forcing to another page to load]
> 
> "Refresh" is not a valid HTTP header, and there is currently
> no formal proposal to make it one.
> 
> While HTTP does ask that recipients ignore headers they don't
> understand, it results in poor interoperability.
> 
> Should [sic] there be a requirement that HTTP servers SHOULD NOT
> send non-standard headers?
> 
> Larry

Non-standard headers MUST be ignored by the client (browser).
(if this isn't in a document somewhere, it should be by now)

"SHOULD NOT send non-standard headers" inhibits 
incremental deployment and development of new headers,
and thus new protocol features.

I'm in favor of permitting extensibility, even by non-standard
headers, provided: 

	HTTP servers MAY send non-standard headers.

	HTTP servers MUST NOT require any particular
	response to a non-standard header.

Joe