A question about the standard

[William Pence]

Giuseppe Vacanti wrote:

> Now my question is: is it legal to have (before the END key word) a
> card image filled with blanks? 
>
> I have this problem in the header of a binary table extension. Trying
> to answer this question I noted that FTOOLS's fverify says the file is
> ok, FTOOLS's fv cannot display the extension containing such cards
> (but it doesn't complain), CFITSIO reports an error (202=key word does
> not exist), FTOOLS's fdump will gladly show the header. 

Clive Page sent me a FITS file with this problem and it turned out not
to be related to the blank keywords, but instead due to the following
TDISPn keywords:

TDISP3  = 'f8.4    '           / display format 
TDISP7  = 'i8      '           / display format 

The fv program requires that the Fortran display format be given only
in upper case (e.g., 'F8.4' and 'I8') as is implied by the definition
of the TDISPn keyword in the Binary Tables Extension FITS paper.

We plan to change fv to be more lenient and allow lowercase format
characters in the TDISPn keywords, but the question remains whether
this is really legal.  Fortran-90 (and Fortran-77) allow lowercase
format descriptors, but the Binary Table definition document
specifically lists only uppercase format characters.

What say other FITS experts on whether lowercase TDISPn values are
legal?

-Bill Pence
NASA/GSFC

p.s.  I was unable to reproduce the error=202 in CFITSIO, so I suspect
that it is unrelated to this TDISP issue.