decision tree for civil time

From: Rob Seaman <>
Date: Sat, 13 Aug 2005 09:46:37 -0700

This is a first draft of what is intended to be a complete inventory
of options for civil timekeeping. Just compile the options now -
we'll undoubtedly express our opinions about each later.

Speaking of which, newcomers may be unaware of the list archives at There are a lot of
creative ideas in there worth additional exploration. Two that
jumped out at me last night are Mark Calabretta's daily epsilon
corrections and Ed Davies' idea to issue a leap second whenever
possible (i.e., leap 1s forward one month, 1s backward the next,
whenever |DUT1| < 0.9s allows).

What I like about the latter is that a synchronization step is
therefore indicated by the ABSENCE of a leap second. What I like
about the former is that it addresses the actual issue head-on - kind
of refreshing actually. The value of such suggestions is not only
that they could each be used to solve the problem in front of us
(Davies' idea is valid under the current ITU standard!) - but that
they also bring needed clarity to the analysis of the problem.

Please remark on any gaps you see, especially missing categories:

I) Existence
     A) adopt a civil time standard
     B) no civil time standard

II) Multiplicity
     A) a single international civil time standard
     B) multiple civil time standards

III) Locale
     A) restrict focus to Earth
     B) Solar system-wide solution

IV) Synchronization
     A) Type
         1) Solar
         2) Atomic
         3) Other (e.g., sidereal, lunar, dynamical, relativistic, ...)

     B) Overloading
         1) None
         2) Dual (e.g., UTC is synchronized to both Solar and Atomic
         3) Multiple (e.g., Earth+Mars+Atomic)

     C) Method
         1) steps (I'm tired of the word "leap)
             a) arbitrary size
             b) fixed size
                 i) second
                 ii) minute
                 iii) hour
                 iv) other
         2) rates ("rubber seconds")
         3) both

     D) Triggering
         1) manual (e.g., reset a clock by hand)
         2) automatic

     E) Adjustment
         1) discontinuity
         2) time slice (e.g., UTS)

     F) Scheduling
         1) as needed
         2) fixed in advance

     G) Optimization
         1) remain within tolerance (e.g., 0.9s)
         2) minimize figure of merit (e.g., DUT1)

     H) Frequency
         1) continuously
         2) daily
         3) monthly
         4) annually
         5) decade
         6) century
         7) greater than a century

     I) Notification
         1) monthly
         2) six months
         3) annually
         4) decade
         5) fifty years
         6) surprise

I wouldn't be surprised if my biases are still evident. Please help
me to remove these and to avoid adding any of your own. The
structure is also too flat - categories might benefit from
hierarchical nesting.

Rob Seaman
National Optical Astronomy Observatory
Received on Sat Aug 13 2005 - 09:46:45 PDT

This archive was generated by hypermail 2.3.0 : Sat Sep 04 2010 - 09:44:54 PDT