IP address within user site network is used as a component to identify a session, which is the role of IDENTIFIER. Of course, it also plays a role of LOCATOR in user site network. But in the transit network, the IP address mentioned above only has a role of IDENTIFIER.
firstname.lastname@example.org [mailto:email@example.com] 代表 HeinerHummel@aol.com
In my view the pair "DNS name / IPv4 address" is a true ( identifier / locator ) -pair, whereas EID and RLOC are rather two locators separated by some strict or loose hop.
Imho LISP is pure two loose hops routing. Isn't it ? LISP wouldn't make sense if there were no hop between EID and RLOC, would it ? (in that case, what would be the meaning of EID resp. RLOC ?)
So working out the double nature "identifier/locator" is first of all creating a special aura of research.
Is this too critical ?