Summary

  • John Houlker’s official .nz delegation passed to ISOCNZ in 1996, but the 2004 AGM record says IANA’s database was only then correct for the first time since 1996.
  • That lag separates a working registry from the administrative record describing its manager. The minutes do not show a DNS outage, a bad lookup, or which IANA fields were stale.

A handover can be remembered as one date and recorded as another. In the .nz case, InternetNZ’s retrospective places the transfer of domain-space management to the Internet Society of New Zealand in 1996. Yet minutes of the Society’s 30 July 2004 annual meeting say the IANA database for .nz had become correct “for the first time since 1996.” The meeting also thanked the University of Waikato for its support and patience while IANA details were updated. The episode leaves a narrow but useful question: what continued when the administrator changed, and what did the record fail to keep in step?

The sources separate several milestones. IANA’s current .NZ delegation page lists 19 January 1987 as the registration date and InternetNZ as the current manager. InternetNZ’s biography of John Houlker says he held the official delegation from 1989 until it passed to the Society, then ISOCNZ, in 1996. Those dates are not necessarily competing claims: a root-zone registration date and a later account of who officially held the delegation answer different questions. The available record does not establish precisely what happened in 1987 or why Houlker’s formal period is described as beginning in 1989.

A careful history should preserve the distinction instead of compressing it into a single “launch” date.

The institution that received the work was also new. InternetNZ’s 2025–26 annual report says the Internet Society of New Zealand was formed in 1995 and took over management of the .nz domain name space in 1996. The report recalls only a few hundred .nz names at the time, compared with 100,000 by 2001. This was a transfer into a community organization while the namespace was beginning a steep growth curve. But the 2004 AGM record adds an administrative afterlife: the organization’s own meeting still found IANA’s data wrong years after the transfer it described.

The IANA database reference must be handled carefully. The minutes do not identify the stale field, specify the date it was corrected, or say that users could not resolve .nz names. They do not demonstrate a broken root-zone delegation or an interruption to registrations. “The database was not correct” is a statement about a record; the sources supplied here do not map that statement to an outage or a failed DNS query. The motion’s thanks to Waikato and the wording about updated IANA details suggest that legacy contact or management information was involved, but that remains an inference, not a disclosed field-level audit.

The distinction matters because “the registry” is often used for several systems at once. RFC 1591 describes IANA’s coordination of top-level delegations and the designated manager’s responsibility for names and nameservers. It also says that a manager acts as a trustee serving the community, not as the owner of a domain; for a transfer, the old and new organizations should communicate their agreement and the new manager’s understanding of its duties to IANA. The memo supplies a framework for understanding the episode, not a New Zealand transaction file.

The 2004 minutes do not tell us what messages were exchanged in 1996, whether the transfer documentation was complete, or how any delay arose.

Nor was the Internet connection itself the same thing as the .nz delegation. Houlker’s InternetNZ biography separately recounts his work on NZGate and the 1989 link to Hawaii, and says that Internet connection services shifted to telecommunications providers from 1995. A route carrying packets, a nameserver answering for .nz, a registry assigning names, and an IANA record naming the responsible contact have dependencies, but they are not one service. A change in the administrator entry does not itself move the cable; a change in the cable provider does not itself reassign the country-code delegation.

That makes the .nz chronology an unusually clear test of institutional continuity. The administrator changed in 1996, while the organization later recorded a need to repair the upstream account of that arrangement. The record does not let us grade the system as either continuous or broken. It does show that continuity has an administrative layer that can fall behind the people and organizations doing the work. A domain may be operated through live nameservers while its contact record lags; conversely, a perfect record cannot by itself prove the nameservers are healthy.

Heng Lu’s Note 70 offers a present-day lens: protect the registry function without equating that function with the permanence of its current administrator. It is not evidence about .nz. Here, the lesson is modest. The 1996 transfer matters because service responsibility moved; the 2004 correction matters because the ledger describing responsibility needed maintenance. Neither date proves who “owned” .nz. RFC 1591 explicitly rejects ownership as the useful frame for a delegated domain and emphasizes duties and service.

The record therefore supports neither a seamless-handover legend nor a breakdown story. It gives us a transfer reported for 1996, a correction reported in 2004, and an unresolved question about the exact IANA data between them. The useful history sits in that gap: operational continuity and documentary accuracy are related obligations, but they are not interchangeable evidence.

Source records: IANA .NZ delegation record; InternetNZ Fellows: John Houlker; InternetNZ Annual Report 2025–26; ISOCNZ 2004 AGM papers; RFC 1591; current .nz Rules. The interpretive question comes from Heng Lu’s Note 70, not from the historical record.