[fitsbits] Re: generic coordinate system

[David Berry]

> I'm evaluating the possibility to use the FITS format to save the results of
> my numerical simulations. The problem is that I may have not regular grids,
> so
> if for example I save the density rho(i,j) in a plane, the trasformation
> between
> numerical coordinates i,j and physical coordinate system x(i,j) y(i,j) is
> not
> trivial. I'm thinking to save also x(i,j) and y(i,j) as 'images' in the FITS
> file, but
> it would be useful to have a software to display rho vs (x,y).
> 
> Do you know if there is some software that can do it?

You may be interested in the Starlink AST library:

http://www.starlink.ac.uk/star/docs/sun211.htx/sun211.html

This provides facilities for storing, managing and plotting fairly
general WCS schemes, including various pre-defined forms (such as the
proposed FITS WCS), arbitrary algebraic transformations, look-up table
based transformations, etc. AST can stored descriptions of all these
WCS systems within FITS headers using its own keyword system (it can
also read and write WCS in the form of the Greisen and Calabretta FITS
papers proposals, so long as your WCS can be described in terms of the
model proposed by the papers, which may not be the case for yours).

AST also provides facilities for plotting arbitrary non-linear axes
which are independant of graphics system. Facilities for using PGPLOT
as the underlying graphics system are included with the library, but
if you want to use some other graphics package you just have to write
a simple module which tells AST how to draw a straight line and  plot
a character string. Starlink can provide such modules to allow Tcl/Tk
or Java Swing to do the drawing (a JNI binding for the library is also
available, also a fortran 77 binding).

David