I am looking for FITS files where BITPIX = 32 and NAXIS = 2

[Don Wells]

>>>>> "Juan" == Juan Jose Garcia Adeva <jgadeva at jet.uk> writes:
    Juan> Hello, I am using the FITS format in a project, but I only
    Juan> need to support files with BITPIX = 32 and NAXIS = 2. .. I
    Juan> cannot find any example in
    Juan> http://www.cv.nrao.edu/fits/data/samples. ..

Try your code on this 262_KB file:

    http://www.cv.nrao.edu/fits/data/tests/incunabula/m87-32.fits

Examine the two text files m87-32.list and m87-32.vf in the same
directory. Note that ORIGIN='KPNO' and DATE='13/09/79': this is the
oldest FITS file with BITPIX=32. The 'vf' file is the output of a FITS
file verifier when reading this file. It asserts that the '/' in the
keyword 'IPPS-B/P' is illegal, however this file is readable by modern
software (Cotton's FITSview code displays the file properly). The
'IPPS' refers to KPNO's Interactive Picture Processing System, the
predecessor of IRAF and 'B/P' is bits per pixel; the IPPS executed on
a CDC 60-bit computer, and floating point pixels were carried in half
words, probably with 20-bit mantissas but I can no longer remember
that detail. The number system of this FITS test file was specially
contrived: BSCALE=1.00000E-06 was used in order to make the
significant bits be spread across the second and third bytes to expose
byte-flip bugs.  Furthermore, the file asserts that BLANK=-2147483648,
so this header is sensitive to differences in implementations of number
conversion routines, the situation discussed by John Davis and Bill
Pence in the unsigned-integer thread in the past few days.

-Don
-- 
  Donald C. Wells         Associate Scientist         dwells at nrao.edu
		    http://www.cv.nrao.edu/~dwells
  National Radio Astronomy Observatory                +1-804-296-0277
  520 Edgemont Road,   Charlottesville, Virginia       22903-2475 USA