Re: Year 2000 Problem for TLEs??

Bjoern Gimle (bjorn@tt-tech.se)
Fri, 22 Aug 1997 14:37:07 +0200

Bruno Tilgner mentioned some of the programming aspects of the
millenium bugs. I'd like to expand, summarize, and speculate from
(another) programmer's point of view:

1. Many systems will fail because of inherent bugs in BIOS and
   other firmware programs.

2. Some systems may survive this if they are powered off on Dec.31, 1999

3. Programs may crash or fail because they read only two year digits.

4. Really bad programs (often COBOL) are written to believe '99' or '00'
   means 'end of data file' 

5. Some programs work around this by using four digits internally, and
   assuming a minimum value, like 50 or 57, for 1900+, the rest for
2000+.
   This may be the case for some satellite programs.

6. Most programs that compute differences between dates will crash or
   give incorrect results when they span the centuries, if they don't
   work like (5).
   This would be the case for Jan.2000 predictions from Dec.1999 elsets.

7. Some programs may produce inaccurate results even without spanning
   the centuries, like when converting XYZ positions to alt/az (wrong
   coordinate system epoch) and RA/dec. precession/nutation of
coordinates.

8. Some programs are considered to fail because they produce output with
   two-digit years. It is the responsibility of the user ( or program )
   reading this output to interpret it sensibly !

Without looking into the code, I would speculate that NORAD models could
suffer from (7), at least for the SDP4 model used for nodal periods 
over 225 minutes. I can accept Bruno's statement that there is no
intrinsic fault in the model - it is up to the programmer to feed it
the correct time difference for the propagation. But I think SDP4
computes Moon and Sun effects, and thus need an absolute date.
Although the Sun's position is approximately the same in both centuries,
the Moon's (maybe) quite different.
If this (four-digit) year (or Julian Date?) is the programmers' 
responsibility, we are back to situation (5).

-- 

------------------------------------------------------------
-- bjorn@tt-tech.se       (office)   +46-8-59095783       --
-- b_gimle@algonet.se     (home)     +46-8-7428086        --
-- 59.22371 N, 18.22857 E , 44 m     AND member of :      --
-- http://www.algonet.se/~b_gimle    SeeSat-L             --
------------------------------------------------------------