Summary

  • Terry Manderson's 2006 APNIC proposal would have stopped email updates to Whois records in stages over 12 months after adoption.
  • A revised 18-month plan still failed to win consensus; APNIC records the proposal as withdrawn in August 2007.

The most stubborn part of a registry can be its inbox. In 2006, APNIC members could update Whois objects by sending plain text to APNIC’s automated update mailbox, or use the certificate-secured MyAPNIC site. Manderson's first prop-037 text argued that email made authentication weak, admitted spam and viruses into the update queue, and sometimes lost legitimate automated replies to members' own spam filters. The address book was being changed through a channel that could not reliably distinguish a wanted transaction from mail noise.

Manderson proposed a replacement, not merely a locked door. APNIC would offer automated XML/REST transactions over certificate-secured HTTPS, a test system and command-line examples. The plan gave domain-object email updates four months after adoption, address and ASN objects eight months, and remaining object types twelve. Every deadline depended on adoption; none was an operating cutoff.

The APNIC 22 minutes show where the plan hit its real cost. Participants asked why members who preferred MyAPNIC could not use it while others kept their working email tools. Someone noted that email itself could be secured. Others pointed to automation already under construction and asked whether synchronous updates had to be tied to retiring email. The chair found no consensus for the proposal, though the room agreed to form a working group. It was a vote for more engineering, not for Manderson's deadline.

The revised February 2007 text, credited to the APNIC Secretariat, made that cost visible in its calendar. It proposed a prototype at six months and a complete cutoff at eighteen, with object types moving in stages at months ten and fourteen. Six extra months, a test endpoint and sample tools bought a more practical runway. They did not buy consent. APNIC's history records a second failure to reach consensus at APNIC 23 and withdrawal on 16 August 2007.

The proposal exposed a familiar registry asymmetry: APNIC could design the new write path, but members would have to rewrite the scripts connected to it. The primary record gives no count of those scripts or a measured migration bill. That missing inventory would tell us how large the burden really was. The result is plain enough: a longer timetable did not turn a mandatory switch into an agreed one.

Portrait identity reference: Terry Manderson's named IETF photograph.