Your argument might be that "they had enough time to migrate". True.
Yes, they had, but ICANN/DoC bureaucrats hold them hostages.
That's an issue not directly (technically :-) related to this discussion.
IMHO, it is better to shut it down completely. The DNS behaves fine when a server is down, not when it is running and giving possibly wrong information (as ns.eu.net hosted by RIPE already did for former domains of EUnet-France).
Could someone explain *what the real problem* is, if it exists/ed? I am definitely NOT a DNS expert, but I couldn't verify a "wrong" answer returned by ns.eu.net yet.
Otherwise, nsd is a very fine software :-) and I'm glad to see it used in production. Why not trying it on NS.RIPE.NET?
But _if_ there is a problem (and not simply a _different_ behaviour than bind would respond, due to an inconsistent configuration somewhere), then everybody would be better off to get it documented and fixed, I suppose? Wilfried. PS: I am very interested to get this sorted out, because we already have that creature doing business here, and I'm considering to use it for another service soon. Thanks....