TO:  All GCN Sites                                20 Nov 05
RE:  Correction in Sun/Moon calculations

OLD and NEW WAY:
Up till now, the GCN program has been using the TJD and Sec-of-Day
for the burst when calculating the RA,Dec of the Sun & Moon and
for the Moon illumination.  This worked fine in the early days of GCN
when ALL the messages were created in real-time (ie received and
distributed with 1-30 sec after the burst, depending on the source type),
but as time progressed  messages started to occur that were delayed in creation
(ie hours after the original burst).  During this delay,
the Sun and the Moon move (especially the Moon), so the Sun_distance,
Moon_distance, and Moon_illumination change from those values
at the time of the burst.  Now the program uses the values
at the time of issuing the notice (not the time of the burst)
no matter what the delay (seconds, minutes, hours, or even days).

WHICH NOTICES WERE AFFECTED:
The Notices that have substantial delay between the 'burst time'
and the 'issuance time' are the <any_mission>-Retractions and
the HETE Ground_Analysis Notices.  The delays are typically 1-6 hours.
The INTEGRAL TEST Notices also suffer this problem because they are issued
with a fixed TJD and SOD, so as the year progresses, the delay can become
up to months.  This is not so much of a problem, because they are Test Notices
(not real burst notices), but people may get confused if they look closely
at the Sun/Moon RA,Dec positions and the Moon Illumination (eg 6 months
makes a big difference in the Sun position).

ACKNOWLEDGEMENT:
I thank Arne Henden for pointing out the descrepancy in the Sun/Moon positions
in these delayed Notices.

Sincerely,
Scott Barthelmy