[fitsbits] further reopening of Public Comment Period on the CONTINUE convention]
William Pence
William.Pence at nasa.gov
Thu Apr 21 12:24:43 EDT 2016
Hi Walter,
On 4/21/2016 5:00 AM, Walter Jaffe wrote:
> A peculiarity of FITS keywords is that the order of the keywords in a header has no
> semantic meaning with the possible exception of SIMPLE=TRUE. This indifference to
> order often causes some difficulties, but having been established
> as a practice, I think that it is a bad idea to introduce a new variant where
> the exact order of the CONTINUE statements is critical.
The FITS standard is not indifferent to keyword order and states (in
section 4.1.1.1): "It is recommended that the order of the keywords in
FITS files be preserved during data processing operations...".
Thus, it would be bad practice to not preserve the exact order of all
the keywords. In the 20+ years that the HEASARC and others have been
using the CONTINUE convention in huge numbers of FITS files, I am not
aware of a single case of the keywords becoming mis-ordered.
>
> I recognize the importance of being able to specify long strings but prefer
> either ad hoc solutions like those used to specify vectors:
>
> SVALUE1 = ' first part of long string '
> SVALUE2 = ' next part of long string '
> ...
>
> or puting these strings in an extension as an ascii or binary table.
Users can of course continue to use ad hoc solutions like this, but I
hope you can still be supportive of projects that need a more robust
solution for reading, writing, updating, and deleting long character
strings in FITS headers. Incidentally, the HEASARC required that long
string values be assigned to keywords that already had an index number
appended to the name, so your ad hoc solution of using index numbers to
order the continuation keywords would not work for us.
regards,
Bill
--
____________________________________________________________________
Dr. William Pence Astrophysicist William.Pence at nasa.gov
NASA/GSFC Code 662 [Emeritus] +1-301-286-4599 (voice)
Greenbelt MD 20771 +1-301-286-1684 (fax)
More information about the fitsbits
mailing list