Andy, Gavin, Jasdip, Tom, It was good to see -05 of the referrals draft come out with the rdap-related relations and noRedirects0, that is the split we worked out on this thread over the summer, and it made my own document a lot simpler. I have just posted -02 of the subordinate referrals draft to sit on top of it:
https://datatracker.ietf.org/doc/draft-ranjbar-regext-rdap-subordinate-referrals/ The registry still redirects for transport, as Tom held me to, the holder links back with rdap-related-up to the covering object itself, and noRedirects0 is how a client keeps the registry's own record. The rdap-base family is carried only as a holding position, with Jasdip's change to the undirected definition. If the group would rather have the five in rdap-extensions, which I think is the better home, they move there and this becomes just the holder-designated case. Two things I would value the group's view on. Andy, my question from September still stands: can extensions still take the base relations or should we wait for it to turn over? And one that came out of review: rdap-extensions defines a redirect loop as any redirect back to a server already visited during the same query, which rules out a holder sending a lookup for an address it holds nothing for back to the registry's covering object, even though that is a different resource. -02 works around it by having the holder answer with its own record for the covering block. If the intent was to forbid revisiting the same URL rather than the same server, it may be worth saying so in extensions. All the best, K. Kaveh _______________________________________________ regext mailing list -- [email protected] To unsubscribe send an email to [email protected]
