In message <20020822091741.C545@snowdrop.l8s.co.uk>David Laight writes

>I was just pointing out that nanosecond timings are not
>necessary for anything 'normal' that a system would do.

Which only demonstrates that your idae of ``normal'' does not cover
the spectrum of current, or desired, NetBSD users. 

* Lack of nanosecond granularity timekeeping (and Dave Mills'
  "nanokernel" code) is a significant black mark against NetBSD, for
  high-precision NTP users.

* Microsecond timings are already iffy for (for example) correlating
  timestamps across machines wiht multiple gigabit ethernet
  interfaces. 10 Gbit/sec ethernet is already in the lab, and just around
  the corner. A 10Gbit frame can arrive in well under a microsecond.

Microseconds dont really cut it anymore, which is why I suggested its
about time to move to nanonsecond resolution internally.