[LEAPSECS] comments on DRR TF.460-6

Robert Seaman seaman at noao.edu
Mon Sep 20 19:03:38 EDT 2010

Hello Mr. Hanson,

My understanding is that the baroque SG7 process continues with a deadline for comments on 22 September. Apparently such comments are also suitable to be shared with the Department of State. Unfortunately I have not been able to sort out from https://www.ussg7.org/ITAC-R%20Documents.aspx exactly how this is to be accomplished. Thank you for forwarding my comments to DoS. I would appreciate a reply acknowledging receipt of this email.

My highly negative opinion on the notion of redefining UTC remains unchanged:

1) UTC is more than just "the time scale maintained through the General Conference of Weights and Measures". It is specifically a type of "Universal Time", that is, the solar time scale that evolved from Greenwich Mean Time. There is a responsibility here not only to provide a timescale that is acceptable for narrowly interpreted radiocommunication purposes, but that has a specific interpretation and utility for historical timekeeping.

2) Leap seconds are a means to an end, namely a way to synchronize clock time with mean solar time. By all means discuss other alternative mechanisms for a functional mean solar time civil timescale. This proposal permits no such discussion to take place.

3) That the IAU and AAS have taken no formal position on this issue is not equivalent to their stating support. Rather, as a member of both organizations I can state that it most definitely represents an absence of detailed study of the implications. Astronomers and observatories will certainly express a negative consensus when they realize the large cost to astronomy of updating manifold software systems simply to permit a leap-second-less UTC timescale to be thrust upon us. I was active in the Y2K remediation efforts in our community and at my observatory - changing the definition of UTC will be far more disruptive community-wide.

4) That "no other scientific organization or NGO has expressed strong opinions regarding the proposed changes to UTC" is a reflection of a complete lack of understanding of the nature of the change. The universal response of astronomers (let alone the lay public) in my experience has been incomprehension that the proposal could really be to disconnect civil timekeeping from solar timekeeping. Much effort has been needed to make the concept of the proposal clear. It is simply unacceptable to proceed to pursue this change to UTC when stakeholders are unable to participate in consensus-building.

5) Apparently some unspecified groups at the Department of Defense and NASA issued "statements" supporting the proposed changes. What form did these statements take? How formal a process was followed? Did NASA query astronomers at the Space Telescope Science Institute, for instance? Was there sign-off from all missions and divisions that might be affected?

6) On the other hand the National Science Foundation assumed "no position" as did the Department of Commerce. For this draft to state that "no Federal Agency has expressed opposition to the revisions" is equivalent to saying that none expressed support. By far the likeliest reason for not expressing a position on such an issue is a complete lack of understanding of the implications.

7) To say that "leap seconds were devised to keep the UTC time scale in close alignment with earth time making UTC useful for celestial navigation" is to suggest two unsupported assertions. First that no other requirements for "earth time" exist, and second that UTC is responsive only to the evolving needs of those who used to have a requirement for celestial navigation. Rather, no survey of broad civil timekeeping requirements for "earth time" (mean solar time) has been taken.

8) Surveying a narrow range of highly technical organizations for their timekeeping requirements does not address the broad range of civil timekeeping needs reflected in Coordinated Universal Time.

9) It has been asserted that "leap seconds are increasingly causing difficulties with various positioning, navigation and timing systems". Perhaps. How do these difficulties balance with the issues that will arise from the unprecedented historical change that is being contemplated of relayering the entire world-wide civil timekeeping system on a non-solar timescale? Simply asserting there are no such issues should not be deemed sufficient.

10) There appears to be no coherent system engineering management plan associated with this proposed change. What are the trade-offs at work? Asserting a lack of problems (except to astronomers) doesn't make this the case.

11) No alternative possibilities appear to have ever been on the table for discussion. Why is this notion of ceasing leap seconds the only proposal that has been put forward? For instance, the 6 month scheduling window of leap seconds could be lengthened. Any reasonable engineering trade-off study would evaluate several alternatives - including the status quo.

12) Considering Annex 1, question 1 - The linkage between UT1 and UTC is rather between both and a general notion of "Universal Time". Advocating such a change is the same as advocating the destruction of "UT" as a useful concept. There are two flavors of timescale (with many specific implementations): "atomic" time and "solar" time. Civil time is solar time because our clocks are subdivisions of our calendars. Pretending that there is only one type of timekeeping does not make it so. My answer to the question, "Do you support maintaining the current arrangement of linking UT1 and UTC?", is "YES".

13) Annex 1, question 2 - Also ask what difficulties will be encountered when DUT1 exceeds 0.9s. Much astronomical software will have to be rewritten - that is absolutely certain. What other systems will break? Nobody has performed this inventory. My answer to "Do you have any technical difficulty with leap seconds now?" is "NO".

14) Annex 1, question 3 - I urge you to support a different process than has been followed over the past decade or so. UTC is a broader issue than "ITU-R TF.460-6". Billions use UTC. Hundreds have ever heard of ITU-R TF.460-6. My answer to "Would you support the revision?" is "NO".

15) Annex 1, question 4 - Eliminating leap seconds will cost astronomers many millions of dollars merely to preserve current capabilities. Undoubtedly we will be fixing newly introduced bugs for many years. More generally, focusing on "technical difficulties" is only a small aspect of the full system engineering issues at work with the broader issue of civil timekeeping. This is not merely a "precision timekeeping" question. Please reach out assiduously to other stakeholders. My answer to "Would the revision create technical difficulties?" is clearly "YES".

Thank you for the opportunity to comment.

Rob Seaman
National Optical Astronomy Observatory
Tucson, AZ

More information about the LEAPSECS mailing list