Memo:      Calc 9.1. Undocumented features.
===========================================
1999.11.22                                              1999.11.22-18:34:28

1) Constant of precession is modified as: 
     Delta_psi = -0.2957d0  (arc-second/century)
     Delta_eps = -0.0227d0  (arc-second/century)

   When so-called IERS 1996 nutation is computed
   ( ../calc91/cnutm.f )
   Maybe these values fit observations better but they don't comply IERS 1996
   nutation model.

2) Permanent nutation offsets: Delta_psi = - 43.1d0 mas, 
   Delta_eps = - 5.1d0 mas are added when so-called nuataiuon IERTS 1996 is 
   computed. 

   ( ../calc91/cnutm.f,  routine KSV_1996_3 )

   IERS 1996 model doesn't contain these corrections, although, of course,
   they are desirabl.

3) Pole tide contribution is computed incorrectly:
   modeled pole positions are used instead of pole coordinates used as a priori:
 
                    X-model(arc-sec) = +0.0039 + 0.00062*T  
                    Y-model(arc-sec) = +0.1485 + 0.00333*T  
                    where T = number of years since 1945.85.

   It is a violation of IERS 1996 standards. Consequences are:
   a) station positions changed up to 10mm;
   b) station positions are not comparable with another software systems.
      F.e. Bernesse GPS software uses 

	Xmodel = 0.033 arcsec
        Ymodel = 0.331 arcsec
   
   c) the notion of field of velocites is revised. If two stations on the
      surface doesn't have a relative motion with respect to each other 
      (f.e. measurement foppler effect reveals v=0) then SOLVE will yeald some
      small relative velocity and vice versus. The effect of changes of station
      coordinates due to secular pole motion is and observatble effect, 
      it doesn't depend on coordinate transformation. Removal of the effect
      which can be observed corrupts our velocity field. SOLVE doesn't have the
      way to resotre true velocity.