[daip] CLCOR SU table update problem (?)

[Leonia Kogan]

Hi Lincoln,

I asked John Benson to repeat the calculation using CALC.
But as I understand it is not simple to do it, because CALC does not 
carry out the direct recalculation of the coordinates and John needs 
to design a special tool to carry out the recalculation.
I am affraid that an additional mistake can be added at the tool creation.

I have carried out two more tests comparing JPRECS and SCHED for diferent times and sources. I got the differenses less than 0.6 mas. This number is under 
1mas which is (as Craig supposes) the SCHED's accuracy

The software used by SCHED is independent on JPRECS. So the small difference 
of SCHED-JPRECS at the all three tests confirm (I think) the accuracy of 
JPRECS at the level <1mas.

I am enclosing the CRAIG's answer on my request.

Leonia 
----- Begin Included Message -----

>From cwalker at zia.aoc.NRAO.EDU Mon Oct 22 13:36 MDT 2001
Date: Mon, 22 Oct 2001 13:36:57 -0600 (MDT)
From: Craig Walker <cwalker at aoc.nrao.edu>
To: lkogan at zia.aoc.NRAO.EDU
Subject: Re: [daip] CLCOR  SU table update  problem (?)
X-Sun-Charset: US-ASCII
Content-Type: text
Content-Length: 1406
X-Lines: 27
Status: RO

I don't know what the expected accuracy of SCHED's transformations is.
You can look at the SLA code, if you want, but I don't see accuracy
statements there.  If you look at ~cwalker/sched/src/Sched/schpre.f to
see which routines are called, then look in ~cwalker/sched/src/Sla for
the actual routines.  I think these are meant to be implementations
of the "official" algorithms.

I found a USNO web site that offers a subroutine package (in either
Fortran or C) that claims to be identical to what is used to produce
the Astronomical Almanac (which is produced by USNO).  It claims
precision to better than 1 mas.  You seem to be at the 0.25 mas level,
so it's not clear how much agreement to expect at that level.  You
might find the introductory comments to the page informative.  It is at
http://aa.usno.navy.mil/software/novas.novas_info.html.

Also see the SOFA (Standards of Fundamental Astronomy - an IAU initiative)
site at http://www.iau-sofa.rl.ac.uk.  While this makes no claim to
be the Sla library, looking at a couple of example routines, I am
very suspicious that it is derived from the Slalib.  The typing style
is the same and both are from the Rutherford Appleton Laboratory.
Perhaps I should convert to it in SCHED at some point.  But not now.

It would be interesting to know what all CALC does compared to the above.
But your best bet there might be to ask the geodesy folks.

Craig
 


----- End Included Message -----