nodeID

Dear wg,

maybe i'm missing something: Is there mentioning  rdf:nodeID is not 
supposed to survive parsing at another place than in the syntax and 
model document?  

Isn't it rdf:nodeID's provide a simple hack, which allows to (re)-serialize
RDF  graphs as  rather flat verbose XML (with literal parsetype as 
exception)?

I would like to see cases -if there are- described in the primer, which 
require (!) the use of rdf:ID, rdf:resource and other syntax constructs -

rs