On 2/19/26 12:27 PM, Jasdip Singh wrote: > As for the RIR Search relation types, it says: > > "Furthermore, servers SHOULD only use the "/redirects0_ref" path in an HTTP > redirect when the link relationship type is one for a terminal relationship > such as "rdap-top" and "rdap-bottom" (i.e., "rdap-up" and "rdap-down" do not > explicitly express a relationship that is the end of a series of redirects).” > > I think beside the termination criterion here, another criterion in returning > a single object could also be considered. Since “rdap-up” and “rdap-top” > searches could both at most return one object whereas “rdap-down” and > “rdap-bottom” searches could return zero or more, including one sometimes. > Therefore, using "the "/redirects0_ref” path in an HTTP redirect should be > possible for all these relation types, it seems. >
This sentence is about being redirected to a terminal point, not about being redirect to a lookup or a search. Maybe it is misguided or should be soften down from a SHOULD, but it seems odd to respond to rdap-bootstrap with an rdap-down or rdap-up. That basically means that the client gets a redirect to a redirect to a response that may have links that it needs to dereference to get to the answer. Or maybe this should be removed as server operators might know best what their users want. > Also, could expound that IANA RDAP Bootstrap files would most likely be > employed to determine such RIR1-to-RIR2 redirections. Or the RIR extended stats files to short circuit even more redirects due to intra-RIR resource transfers. > > For these relation types, might help to informatively refer to RFC 9910. Agreed. -andy _______________________________________________ regext mailing list -- [email protected] To unsubscribe send an email to [email protected]
