Re: Seeking clarifaction of tai64nlocal and leap seconds

"Andy Bradford" <[email protected]>
writes:

> Nowhere does he claim that it will eliminate the need for a GPS source
> or other time sources. He only claims that his software will eliminate
> clock skew. But, after 3 months of no network time input, 0.21 seconds
> off seems pretty impressive to me.

It all depends on what your goal is.  ntpd will generally maintain the
clock considerably more accurately than that, but you may not need the
additional precision.

0.21 seconds off is enough to occasionally cause weird build problems if
you're running make on separate hosts with a shared network file system
that has sub-second timestamp precision.  Many people don't have this
problem, of course.  :)

-- 
Russ Allbery ([email protected])             <http://www.eyrie.org/~eagle/>