[LEAPSECS] Leap second mess at NASA

[Robert M. Candey]

Thank you all for your comments on my "Requirements for handling leap 
seconds in CDF" page.  As you know, describing time consistently and 
accurately is difficult.  My document was written to provide 
background to how we are developing a new time variable type for use 
in the Common Data Format (CDF) scientific format used by the 
heliophysics community, and is not intended as a formal position on 
time systems and usage.

Regarding section 3.3, each spacecraft mission has unique constraints 
on their internal time keeping due to onboard hardware and processes. 
My list of how some missions handle time internally was to show the 
variations in use; the data are converted to UTC for external use.

When CDF was developed some years ago (around 1984), the EPOCH time 
variable type was defined as milliseconds from 0AD with fixed 
24*60*60 seconds per day for simplicity, with no specified time scale 
or provision for leap seconds; hence our recent requirement for a new 
time variable type and metadata.  The new variable type is still 
being evaluated by the CDF user community, and your comments are 
welcome as well.  Thank you already to Steve Allen for suggestions 
for my web page.

-- 
Robert.M.Candey at nasa.gov
NASA Goddard Space Flight Center, Code 672
Greenbelt MD 20771 USA

The contents of this message are mine personally and do not reflect 
any position of the US Government or NASA.