No it wasn't! NSEC3 is crackable the same way a 1990s Unix password file is. This was such a big problem that two competing approaches were introduced to defeat it: "whitelies", which I perceive as the "best practices standard" answer, requires servers to operate as online signers (they should have been all along) so they can generate dynamic chaff records to foil enumeration, and NSEC5.