W3C home > Mailing lists > Public > www-html@w3.org > June 2005

Re: xhtml2 question: http-equiv

From: David Woolley <david@djwhome.demon.co.uk>
Date: Fri, 3 Jun 2005 21:02:31 +0100 (BST)
Message-Id: <200506032002.j53K2VF16223@djwhome.demon.co.uk>
To: www-html@w3.org

Whoops sent long reply direct, rather than to list.  I'll 
summarise to the list anyway.

> Sorry if this is a FAQ, but I was wondering how would one do
> refresh/redirect and all sorts of htt-equiv in XTHML2?

http-equiv was intended to supply information to the server, not
the browser, so isn't needed in the to-browser protocol.  In practice
it isn't used by the server.

The HTML 4 specification identifies the use of refresh for redirects
as an abuse, but doesn't formally define refresh.

Refresh is accepted by the big 2 as a real HTTP header (not just in
meta and not just on HTML), but is a proprietary extension under the
marketing name "server push"; it's for slide shows.

Refresh has accessibility implications.  You must not flash up a page without
allowing a slow reader enough time to read it.

Redirects should be done with 30x statuses.  They work for all browsers, and
anyone with enough access to generate server side dynamic pages can send them.
Anyone authoring even static commercial pages ought to have the necessary 
access.

> If yes, what's the namespace containing refresh and all the HTTP
> properties (like cache-control)?

cache-control would only work if servers supported meta http-equiv.  In
general shared caches only look at real headers (layering principles) and
browsers are allowed to cache more aggressively than instructed to do.
cache-control and pragma are often used with options that are only meaningful
in the client to server direction, and always to defeat a good feature of
HTTP/1.1.
Received on Friday, 3 June 2005 20:08:24 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 5 February 2014 07:19:03 UTC