- From: Charles McCathieNevile <charles@sidar.org>
- Date: Fri, 30 Jan 2004 14:16:05 +0100
- To: <Kurt_Mattes@bankone.com>
- Cc: <david@djwhome.demon.co.uk>, <w3c-wai-ig@w3.org>
On 30 Jan 2004, at 13:44, <Kurt_Mattes@bankone.com> wrote: >>>> but the server must return a 403 or 404 status for the broken URL, > > Some reason the server cannot return a page with page not found > information? None at all. And any decent server will be configured to do this. As the content of a 404 message. A semantic web - one where information can be semantically linked to other information across the web - appears to provide more accessibility than a collection of documents that can point to each other. This should be no surprise - we know that a document where individual pieces of information can be identified and described is generally more accessible than one which simply provides a presentation of a "whole document". Think about how flash and pdf accessibility work, or why HTML is generally preferred over un-marked-up plain text. Think about the ease of manipulating word documents that have style sheets, and the hassle if they just have formatting tags applied. (Hint: The strategy used for those that just have formatting is to work out search routines that assume each set of format rules is effectively a style rule). So what makes a semantic web work? One of the important things is that you can find information on the web - it doesn't move around randomly. Another is that documents respect the semantics inherent in HTML. If a tool thinks it is getting a real page back, it should try to treat it as a page. If it's getting a special instruction then it expects that to come according to the standards, so it can deal with it. For example, a WWAAC system to search the Web for symbols and illustrations to represent a page can also search for known simpler representations. If it goes to a page and is told the page isn't there with a 404 response, it can look to see if there is still a simpler version of that content around, and present it. If it gets a response saying you should try and second-guess the way that the system has organised their documents and mess around with the URI, as an ordinary page, it will try to present that information. For people like the clients of peepo, who probably only followed images to get to the site (hands up if you tried to read an arabic news site in the last year by following pictures?), this information is next to useless anyway. Whereas if the software finds nothing, it can present a "we don't know where this is anymore" that is appropriate for the user. >>>> Returning anything except an error response breaks link checking > software. > What does link checking software have to do with user accessibility? Cheers Chaals -- Charles McCathieNevile Fundación Sidar charles@sidar.org http://www.sidar.org
Received on Friday, 30 January 2004 08:18:25 UTC