Re: [ACTION] represenation of simple and more complicated HTTP conversation in RDF

Shadi Abou-Zahra schrieb:
> To elaborate a little more on this, this is how I understood the proposal:
> 
> Scenario 1: Content negotiation
> - The first HEAD request to a URL yields a response that contains some 
> header information; this is wc1. A subsequent GET request with specific 
> parameters yields the actual content; this is wc2. Since wc2 is what has 
> actually been tested, wc1 could be dropped. The earl:subject points to 
> the wc2 instance.

The first request will also be a GET request, since the user agent does 
not know that the server will response with a 300. Whether wc1 could be 
dropped depends on whether you perform any tests on the response.

> Scenario 2: Compound content
> - The first GET request to a URL yields a response with actual content 
> such as an HTML document; this is wc1. Subsequent GET requests are sent 
> to fetch other parts of the "Web unit, for example the CSS and other 
> files. These are wc2...wcN. The earl:subject points to the collection 
> with wc1...wcN since these have been tested collectively.

Yep

> Scenario 3: Redirect
> - The first GET request to a URL yields a response that contains a 
> redirect; this is wc1. A subsequent GET request (to a different URL) 
> yields the actual content; this is wc2. As in content negotiation, wc1 
> could be dropped (though it may make sense to record it in this case so 
> that the URL change is reasoned).

Whether wc1 could be dropped depends on whether you perform any tests on 
the response. E.g. you can check that the redirect location is also a 
link in the response body, or the location is an absolute URI.

-- 
Johannes Koch - Competence Center BIKA
Fraunhofer Institute for Applied Information Technology (FIT.LIFE)
Schloss Birlinghoven, D-53757 Sankt Augustin, Germany
Phone: +49-2241-142628    Fax: +49-2241-142065

Received on Wednesday, 13 December 2006 14:14:43 UTC