Simple scalable method for detecting broken links...

Hello TAG,

I'm not entirely sure if this is the right place to bring this up, but '...to help coordinate cross-technology architecture developments inside and outside W3C' sounds about right.

I've thought of a simple way to alert sites of when they have a broken link, and to do it without generating thousands of reports about one link to the offending site's webserver. I'll be amazed if no-one else has thought of this, so if they have feel free to tell me to shut up, and if they haven't please give me an idea of where to raise the idea.

Essentially its a simple extension that can be added to any webserver and browser to alert sites about broken links in their content. My idea is based on the 404 HTTP response code, and sending a message via SOAP to a predefined receptor on the original webserver, and maybe extensions to HTTP itself (multiple techs, hence sending it to TAG first). It basically works like this:

1 Browser sends HTTP GET for URI to Webserver 1
2 Webserver 1 sends HTML page
3 User follows link
   3.1 User clicks link pointing at Webserver 2
   3.2 HTTP 404 response - target not found
   3.3 Display 404 page to user as usual
4 Browser sends SOAP fire-and-forget message to Webserver 1

  (Alternatively I suppose it could be done via HTTP extension framework?)

What happens now is that the link is somehow marked as 'Reported Broken' so other browsers do not generate the same alert to the web server. not sure quite how to do this. I suppose a response to a HTTP GET containing a HTTP header that has a list of links that are broken on that page might do the trick. (Better than filtering the output to re-write the link with markup saying it might be broken)

The entire details are attached. Does anyone think this vaguely useful? Am I talking to the right people!?! If not then I apologise and I'll climb back into my hole... 8o)

Regards,

Adam.


Adam Fowler BSc MBCS
E-Systems, LIS
University of Derby