- From: Phil Archer <phil.archer@icra.org>
- Date: Fri, 15 Apr 2005 17:22:56 +0100
- To: "SWIG" <semantic-web@w3.org>
David Powell wrote: > > Quoting Ora Lassila <ora.lassila@nokia.com>: > >> Now that you mention it, I realize I have also experienced some "strange" >> behavior. When the Wilbur [1] HTTP client requests a URL (with Accept: >> header containing only application/rdf+xml), I sometimes don't get any >> response and sometimes get the corresponding text/html response (from >> some >> servers that are able to differentiate between the two) when I am behind >> our >> corporate firewall, but the correct RDF file when I am on the "open" >> internet. I haven't had time to debug this yet, though. > > Is the server sending back a "Vary: Accept" header? If not then your > proxy/firewall might be cacheing the response when you access it with a > browser, and returning the same HTML response when you try to access it > with > your RDF client? > I don't think that was the problem. When the MIME type was set correctly, pointing a browser directly at the RDF/XML instance got no response. The Worrying thing was that browsing the site itself worked intermittently. That is, some pages were accessible in the browser, some weren't. Everything on the server included an HTTP Response Header that pointed to the RDF instance. (As is still the case, see the headers at [1]) But, caching might be an issue somehow. When the system was very first set up I _did_ manage to see the headers for the RDF instance correctly using the same tool pointed to in [1] (run by DJ Delorie). The tool on the ICRA website [2] uses Perl's LWP to make network requests and has an accept header of */*. So I've not been using anything as sophisticated as a dedicated RDF client. Phil [1] http://www.delorie.com/web/headers.cgi?url=http%3A%2F%2Fwww.madesafe.com [2] http://www.icra.org/RDF/label/tester/
Received on Friday, 15 April 2005 16:25:40 UTC