# The following is a draft version of the ACE/ULEIS project 
# binary time standard.
# 
# Known Bugs:
#
# Revisions
# V 1.0 S.Nylund 06 Mar 1997
#
#******************************************************************************

DESCRIPTION:
The ACE Science Data Center is not putting forth a mission-wide standard, but 
will adapt to each team's standard.  The ULEIS team is therefore unburdened by 
a standard but left with choosing one.  Different ideas will be presented with 
comparisons made.

The format of the chosen standard should be a single, ever increasing 
(monatomic) value; alternate formats that use several values (such as year, 
day-of-year and seconds of day) have the continual problem, when increasing 
the highest resolution element, of needing to watching for a roll over in that 
element.  The monatomic value needs to be referenced to some time.  Past 
missions have chosen times like 1 Jan 1960 or 0 A.D.; alternatives include 
the launch year and launch date.

REFERENCE POINT	            COMMENTS	
--------------------------  ---------------------------------------------------
1 Jan 1997(year of launch)  A fairly easy point to remember but a choice that
                            would be unique to ULEIS

1 Jan 1960                  A round reference point of no particular 
                            significance used by other missions.

Date and time of launch     A point that cannot be establish in advance; 
                            another reference would have to be used prior 
                            to launch when testing software.  Again, this is a 
                            standard that would be unique to ULEIS.  It would 
                            have the limited value of always representing 
                            time since launch.

0 A.D.                      A standard used by ISTP although not popular one.
                            The number, a floating point value, is very large,
                            but still maintains millisecond resolution.



Time should be measured either in units smaller that the minimum period that 
needs to be resolved, for example integer milliseconds, or in a some small 
unit that is stored as fractional values (e.g. floating point seconds).

UNITS         COMMENTS	
------------  -----------------------------------------------------------------
Seconds       Requires floating point value to maintain millisecond resolution;
              floating point operations are slightly slower than for integers.
              Does allow for precision less tham a millisecond.

Milliseconds  Could be an integer (maintains millisecond resolution, but no 
              higher) for processing speed but would be .

Microseconds  Probably higher resolution than would be needed.

S/C clock     Uncertain what the resolution is; will be the time standard in 
              the L1 files coming from Cal Tech.



At present the leading candidate is floating point seconds sinc 1 Jan 1997.
The team will finalize a decision.


While a monatomic standard is very well suited for processing, people do
not think in such units.  Conversion routines from the internal standard to
units of date and time in various formats will be needed.