- From: Graham Klyne <GK@ninebynine.org>
- Date: Mon, 04 Oct 2004 17:25:30 +0100
- To: mdirector@iptc.org, www-rdf-interest@w3.org
At 17:07 04/10/04 +0200, Michael Steidl/MDir IPTC wrote: >We intend to create a specific RDF Schema within an RDF system of another >party >- and hence have to identify our resources by our own namespace. > >Looking at the RDF specs and a lot of examples I found virtually all RDF >namespaces are made from the http URI schema and all have as last >character a # >or a /. > >This raises these questions: > >- is there anywhere a written requirement for having only URIs from the http >schema. No, you can use any scheme. But many people argue that it's best to use a form of URI that can be dereferenced to obtain a description of the resource concerned. There's also an ongoing debate about the nature of what http: URIs denote (cf. http://www.w3.org/2001/tag/issues.html#httpRange-14). >- where does this #- or /-sign requirement come from RDF forms URIs from QNames by concatenation of the namespace URI and local name. Having a '/' or '#' at the end of the namespace URI makes it easier to reverse this process. >- Finally: as we usually specify all our XML namespaces with URIs from the >URN >schema would this be possible for a RDF Schema - as URNs don't allow for >having >/- and #-signs. It's true that URN's don't (strictly) allow '/' signs, but they do not prohibit '#' signs, as the fragment is not part of the main URI. See http://www.w3.org/DesignIssues/Model.html for some discussion. You could include an escaped (using %hh) '/' in a URN. #g ------------ Graham Klyne For email: http://www.ninebynine.org/#Contact
Received on Monday, 4 October 2004 16:25:20 UTC