changes.txt

一个预测卫星方位俯仰的软件

Release 2.2.1d:
by John A. Magliacane <[e-mail]> (04-Nov-2002):

* PREDICT's orbital model has been replaced by more precise
  SGP4/SDP4 algorithms for increased tracking and orbital
  prediction accuracy. 

* The Single Satellite Tracking Mode has been re-written to permit
  support for a transponder database.  It is now possible to select
  precise Doppler-corrected uplink and/or downlink frequencies
  interactively when tracking satellites in real-time.  (No more
  hunting around the transponder to find your downlink!)

* Numeric co-processor emulation linked during the compilation phase
  to permit operation on older machines lacking a co-processor (such
  as 80486SX and 80386-based PCs). 

* Several new command-line options added to permit simplified satellite
  location and orbital prediction to be conducted either via the
  command-line, or from within other programs.  This makes it possible,
  for example, for another program to invoke PREDICT to calculate the
  location of a satellite when the timestamp of a CCD image is determined.

* The -u command-line switch has been modified to allow multiple file
  names to be used when updating PREDICT's orbital database from the
  command-line.

* Solar eclipse angle depths, propagation delay, and echo duration are
  now provided by PREDICT.  If ALAT/ALON coordinates are known, squint
  angles are also provided in both the Single Satellite Tracking mode
  and orbital prediction output screens.

* Added serial-port rotator control support for the PIC/TRACK,
  EasyTrak Jr., or other EasyComm2-compatible rotator controllers.
  (A suitable DOS serial port driver is still lacking, however.)

* Main Menu option [I] now provides information on the files loaded
  by PREDICT, the status of the AutoTracking function, the COM port
  used by the tracking interface, etc.

* Compilation was performed with an updated compiler (gcc 3.2).

* The documentation was updated, and many other coding changes were
  made.  See the NEWS.TXT file for additional information.


Release 2.0.5d:
by John A. Magliacane <[e-mail]> (18-Nov-2001):

* Modified the AosHappens() function to prevent a division-by-zero
  error that would occur if used on a satellite containing no valid
  TLE data.  This error caused the MultiTracking mode to crash if
  the predict.tle orbital database contained less than 24 satellites.

* Real-time tracking is no longer performed for satellites that
  appear to have decayed from orbit.

* Fixed an array overflow problem in the Solar Illumination
  Prediction Mode.

* Updated documentation.  Added PDF and HTML versions.

* Some other minor coding changes were made.


Release 2.0.4d:
by John A. Magliacane <[e-mail]> (04-Jan-2001):

* Fixed several small math errors.  One goes back to some very old
  (and widely used) satellite tracking code!  Thanks to Jordi Mas
  for pointing them out and providing corrections.  Thanks also to
  Andreas Trotschel for pointing out an error in the size of the
  global output[] array.

* Portions of the MultiTrack() and SingleTrack() functions were
  re-coded for greater simplicity.

* A bug in parsing the command-line options was fixed.


Release 2.0.3d:
by John A. Magliacane <[e-mail]> (17-Jun-2000):

* Several small bugs in the KepCheck() and SaveTLE() functions of predict.c
  were identified and repaired.  The problems centered on the fact that
  unsigned chars were used to accumulate TLE checksums.  Occasionally,
  the checksums exceeded the capacity of an unsigned char, and rolled
  over through zero as a result.  These variables were changed to
  unsigned integers.

* A small problem in the AutoUpdate() function in predict.c was identified
  and fixed.  Occasionally, Keplerian data for a satellite would not
  update if it had the same reference epoch as that in PREDICT's
  orbital database.  This problem was the result of extremely small
  rounding errors introduced by the complier that caused intermittent
  logic errors when trying to compare two double precision floats for
  equality.  This problem was solved by changing these variables from
  doubles to floats.

* The Keplerian "torture test" was a bit too brutal for some element sets.
  The check for a '-' on position 59 of TLE line 1 was eliminated since
  some element sets have a '+' in that position.


Release 2.0.2d:
by John A. Magliacane <[e-mail]> (01-Apr-2000):

* A solar illumination feature was added to predict the percentage
  of time a satellite spends in sunlight per day.

* Orbital predictions for satellites that appear to have decayed
  since the last Keplerian orbital update are no longer attempted
  by the program.

* Date and time used to start orbital predictions may now be
  abbreviated to Day/Month/Year only (00:00:00 is assumed).
  The default remains the current date and time.

* System clock/calendar is now read to millisecond precision,
  permitting more "lively" real-time tracking mode displays.

* The UTC timezone offset now may be set correctly for users living
  in half hour timezones.

* Internal timezone calculations are no longer confused by environmental
  variable TZ.  If TZ is properly set to a timezone name (as required
  by some other satellite tracking programs), the TZ setting is used by
  PREDICT instead of the UTC timezone offset specified in the predict.qth
  file.

* The command-line parsing code was modified to prevent segmentation
  faults from occurring if command-line options are incorrectly issued.

* Some other minor coding changes were made.