Path: utzoo!mnetor!uunet!husc6!cmcl2!rutgers!sdcsvax!ucbvax!UDEL.EDU!Mills
From: Mills@UDEL.EDU
Newsgroups: comp.protocols.tcp-ip
Subject: Re:  Leaping Clocks
Message-ID: <8712311645.aa28586@Huey.UDEL.EDU>
Date: 31 Dec 87 21:45:01 GMT
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The ARPA Internet
Lines: 25

Phill,

Well, I have good news and bad. The good news is that the umd1.umd.edu
timeteller host will offer rigorous standard time pre-attack, trans-attack
and post-attack the Leap. It is now squawking "danger: leap imminent"
to its Network Time Protocol peers even now as per RFC-958. All fuzzballs
on the U Delaware, U Maryland and Linkabit LANs will automatically follow
the leap as required and relay the leap indicator to their pals.

The bad news is that the wwvb.isi.edu timeteller got its clock punched, or
something like that, since I can't reach out and touch its leap switch. It
will miss the leap, at least until its radio clock notices, and thus keep
NYT for probably about ten minutes while the radio clock resynchronizes.
The remaining timetellers at Ford and NCAR have not had the software update
required to leap correctly.

Having said that, understand the radio clocks themselves have no provision
to trach the leap, so will experience a period of befuddlement immediately
following the Leap of from a few minutes to hours. Thus, my exquisitely
crafted timewarp will come down at the leap, only to be yanked back maybe
a minute later when the disagreement with the radio clock is noticed, then
when the radio clock regains its marbles the unleaped leap will be releaped.
Got that?

Dave