Re: netCDF time representation

• To: netcdfgroup@xxxxxxxxxxxxxxxx
• Subject: Re: netCDF time representation
• From: Steve Emmerson <steve@xxxxxxxxxxxxxxxx>
• Date: Wed, 12 Feb 1992 07:48:55 -0700

>Here are a few examples of data requiring more than 32 bits of precision:
>
>1. Radio pulsar timing         --- 0.1 microseconds over years
>2. Astronomical X-ray data     --- 0.01 microseconds over years
>3. Planet Ephemerides          --- 1 microseconds over decades
>4. Laser/Atomic Spectroscopy   --- 0.1 femtoseconds over milliseconds
>5. Fusion                      --- 1 nanoseconds over seconds
>6. Neutrino detection          --- 1 millisecond over decades
>
>=Fred Knight   (INTERNET:knight@xxxxxxxxxx)

Excellent!  Thank you!

I might add that Mitch Baltuch, a Unidata colleague, informed me that
lightening data requires microsecond resolution over years.

Now, does anyone know of any datasets requiring more than 54* bits of
precision?  This is approximately equivalent to microsecond resolution 
for 600 years.

You'll see why I ask this question in my reply to Daniel Packman's
article about "base arithmetic".

--Steve

* 54 is the number of bits in the mantissa of an IEEE double.