- From: Bob Ferris <zazi@smiy.org>
- Date: Sat, 15 Oct 2011 11:44:50 +0200
- To: public-vocabs@w3.org
Hi, On 10/15/2011 1:11 AM, Gregg Kellogg wrote: > On Oct 14, 2011, at 3:32 PM, John Panzer <jpanzer@google.com > <mailto:jpanzer@google.com>> wrote: > >> Thanks; with regard to this, I wasn't sure what "@itemprop names which >> are not absolute URIs are resolved as relative URIs either to >> @itemtype or Document base." means. > > Basically, if an @itemprop is a bare name, and not an absolute URI, it > is added to the > @itemtype URI using the described mechanism. If there is no @itemtype, > it uses the document base and is added to it using the appropriate > algorithm for resolving relative URIs. > > For example, if @itemprop="name" and > @itemtype="http://schema.org/Person" the resulting URI would be > http://schema.org/name. > > If there is no @itemtype, better to generate something rather than > nothing. So, if base were http://example.com/foo, @itemprop would > resolve to http://example.com/name. > > Of course @itemprop could be something other than a bare name, such as > "foo/bar" and the same rules apply. Following these examples of microdata, I really don't understand why there are such problems with the utilisation of prefixes and CURIEs, which I personally find a bit clearer and easier extensible and mixable than this kind of "wanna be" (not so intuitive) prefix approach. In object-oriented programming languages we usually intuitively work with package names and shortcut (prefix) mechanisms (รก la import statements at the beginning of a class definition). I don't really understand why there are such big problems of applying this mechanism to knowledge representation as well (as it is already done with the prefixing mechanism in RDFa, RDF/XML, Turtle/N3 and SPARQL). Cheers, Bo PS: I guess this is more a topic for the microdata mailing list ;)
Received on Saturday, 15 October 2011 09:45:35 UTC