[dmt314159]

The rationale is that the reference NTP implementation codebase was old and would benefit from a large cleanup. It was in a similar position to openssl in that the code was used by large numbers of commercial entities yet they provided no steady stream of money to maintain it. Openssl only had one paid developer in pre-heartblead days, while NTP only had Harlan whose efforts were maligned a few years ago.

Both of the projects also generally required knowledgeable people as they're in specialist areas.

Like any implementation of NTP, NTPsec needs to work at the protocol level, but I don't know whether it was practical for the NTP project to accept the degree of change in it's codebase that the NTPsec people wanted. In an ideal world the fork could have been avoided and a more vibrant community of developers would have happened.