wcs.ps and scs.ps

[Eric Greisen]

Michael Nolan writes:
 > Comments on the wcs paper (1999 April 9 version):
 > 
 > I can reasonably forsee the usefullness of airmass and unit optical depth
 > as axis dimensions.  

     I am unfamiliar with airmass, does it have a clear definition?
Optical depth is unitless.  Can you describe how specifying that it is
optical depth in the units string (rather than the CTYPEn string) is
needed.

 > 
 > In table 4, adu means A/D converter.  Do you mean converter (as in
 > physical device), or do you mean DN?  I can see a value in both, and
 > have a specific need for the former, since we have four parallel IF
 > chains simultaneously sampled.

    I am also unfamiliar with these units.  Can Francois or one of the
units gurus at OGIP comment on this?

 > 
 > You removed the trig functions, which are mainly useful for projections,

    Actually they have no use whatsoever in celestial projections.
The only units strings are 'deg' in Paper II, which should be
announced in < 1 day I hope.

 > and can be handled in paper II, but have other applications, such
 > as airmass again.  I suspect that if trig functions are needed, they
 > will simply be used withut difficulty.  The main problem would be how to
 > represent the inverse trig functions, as it would be better to choose now than
 > to have all of the possibilities be used later.

     I do not know what use inverse trig functions could have and how
to attach meaning to them as units.  Do you have any specific examples
in mind?  Please describe some in detail.  I agree that cos(elevation)
axes have occurred to me.  Are there other uses for even the forward
trig functions?  We need help from the community in such matters
because our experience has serious limitations.

BTW: I have put a new version of wcs.ps out with a few additions to
the pixel regularization section - a keyword to specify the magnitude
of the correction on each axis.  See
     ftp://ftp.cv.nrao.edu/NRAO-staff/egreisen/wcs.ps.gz


Thanks, 

Eric W. Greisen