Non locator use of URIs

I'm collecting a list of non-locating, identifying uses of URIs. e.g. 
where are URIs used in practice as identifying strings? So far I have:

XML Namespaces
SAX Feature and Properties
Atom IDs
RDF/Owl/SemWeb
XML encryption data types and algorithms
XML canonicalization algorithms


Can anyone add to this list? The criterion is that the URI is not 
necessarily resolvable, and not intended primarily to be resolved. It's 
OK is some can be (e.g. RDDL documents for XML namespaces) as long as it 
doesn't have to be to serve its main function.

-- 
Elliotte Rusty Harold  
XML in a Nutshell 3rd Edition Just Published!
http://www.cafeconleche.org/books/xian3/
http://www.amazon.com/exec/obidos/ISBN=0596007647/cafeaulaitA/ref=nosim