I've previously[1] talked about how stupid #WHOIS is, and while #RDAP is an improvement, it's still really just a bunch of information bits based on (regional) convention.
A human can usually quickly identify e.g., the owning legal entity from inspection of of the data, but good luck doing that programmatically. It's infuriating.
@jschauma Your criticism of RDAP is not clear to me. What could be sent by an information service but "bunch of information bits"? What is the regional convention you speak about?
And about the difficulty to do it programmatically, sure, jCard is painful, but just finding the entry with ""roles": [ "registrant" ] is not to hard for a program.
@bortzmeyer Different RIR's RDAP servers have different information (e.g., for 0.0.0.0, 1.0.0.0/8); the "registrant" role is sometimes present (1.0.0.0/24), sometimes missing (1.0.1.0/24); the vcard information for an entity sometimes points to an individual, sometimes a company, sometimes just an address, sometimes a net handle, etc.
@jschauma I assume the differences are from the different models (APNIC has NIR underneath). Hardly RDAP's fault, just a reflection of the fact that not everyone has the same policies.
@bortzmeyer Yep. And different RIRs may have different requirements around what data is present in what format. So yes, my gripe is not with RDAP the protocol, but with the lack of a forced structure of the data the protocol makes available, I guess.
@jschauma Just for fun, here is a #jq script to extract the registrant (but, as you noticed, the biggest problem is in the lack of consistency of data):
Add comment