- From: Charles McCathieNevile <charles@w3.org>
- Date: Wed, 19 Jun 2002 18:05:53 -0400 (EDT)
- To: David Woolley <david@djwhome.demon.co.uk>
- cc: <w3c-wai-ig@w3.org>
In general I agree with David's assessment - currently most Web designers seem to have very little understanding of HTTP, which is important to content negotiation. And most people still rely on an external provider to host content, although more and more systems have web servers built in and people could host their own content. (This doesn't apply too large sites who generally do control their own systems). Actually W3C does negotiate content-types and languages when there are variants available - for example it is possible to request some documents as html, plain text, or rdf, or for some resources to request them as an image or as rdf... To enable this to work with deep linking is actually not terribly difficult. There are versions of a resource available in a specific type, with a URI, and there is a URI for a resource that can be content-negotiated. For example, http://www.w3.org/Icons/w3c_home is a W3C logo, and might be served as a GIF, a PNG, or in other forms. On the other hand http://www.w3.org/Icons/w3c_home.gif is a GIF version of the W3C logo. Either of these can be bookmarked, but it makes more sense in most cases to bookmark the resource in its general version, not a specific representation of it. (Actually the fact that it ends in .gif just shows the way our server is configured - the relevant URIs could easily be .../jgfkgfjg and .../87otygoybig and the system would still work.) cheers Chaals On Wed, 19 Jun 2002, David Woolley wrote: > I wonder if anyone has tried to use HTTP content negotiation mechanism for > serving different variants according to special needs, e.g. to send a sign I think the problem with content negotiation is that it requires server configuration. Most informational page developers are one technology people, and if it can't be done with a meta element in HTML, it can't be done at all, as HTTP is a foreign technology. Slightly more legitimate reasons are that sources of cheap web space tend not to allow any server configuration, so people learning can't play with it and people an a budget can't afford it. The more technical problems with language negotiation are that it is difficult to do well and maintain cachability (not that that seems to matter to todays designers) and it can be difficult to have a satisfactory way of overriding the negotiated language (without reconfiguring the browser. If one negotiates every page, one needs to reconfigure the browser. If one only negotiates the home page, deep links will not benefit. If one has a negotiation page that redirects into a non-negotiated page, for ever page, it may be difficult to convince people linking to the site to pick up the link to the negotiation page rather than the negotiated page they are viewing. The only real life negotiated pages I have come across are Windows Update and Google (W3C negotiates graphics formats). -- Charles McCathieNevile http://www.w3.org/People/Charles phone: +61 409 134 136 W3C Web Accessibility Initiative http://www.w3.org/WAI fax: +33 4 92 38 78 22 Location: 21 Mitchell street FOOTSCRAY Vic 3011, Australia (or W3C INRIA, Route des Lucioles, BP 93, 06902 Sophia Antipolis Cedex, France)
Received on Wednesday, 19 June 2002 18:05:59 UTC