[fitsbits] Keywords "explicitly declared ... long-string"? {External}

[Tim Jenness]

It has just been pointed out to me that we use a TDOCn header with the documentation for the column (which I think is a local extension maybe?) and the comment being long was an alternative source of information for the user. We will drop the long comment for TTYPEn and continue to use TDOCn.

—
Tim Jenness

On Feb 25, 2025, at 08:02, Tim Jenness <TJenness at lsst.org> wrote:



On Feb 25, 2025, at 07:16, Seaman, Robert Lewis - (rseaman) <rseaman at arizona.edu> wrote:

Hi Gregory,

As folks can infer from my ancient comments, I’m leery of relying on such fiddly features, and I also had nothing to do with whatever wording of either CONTINUE or “type long-string” in the FITS standard.

What I’m wondering is whether Rubin or SPHEREx are really using CONTINUE – specifically, that is, with mandatory or reserved keywords. What is the use case?


The use case is people being verbose in the COMMENT attached to the column name.


TTYPE15 = 'deblend_peakId&'    / ID of the peak in the parent footprint. This is
CONTINUE  '&'                  /  not unique, but the combination of 'parent'and
CONTINUE  '&'                  /  'peakId' should be for all child sources. Top
CONTINUE  ''                   / level blends with no parents have 'peakId=0'

Astropy almost manages to read that back okay but somehow puts the last line of the comment at the start of the comment.
This is an attempt to make the table definition a bit easier to understand if someone is looking at the header. Not sure what our alternative options are though.
All we can think of is adding a `HIERARCH TTYPE15 COMMENT` card explicitly with that comment text in as a long string.

—
Tim Jenness
Vera C. Rubin Observatory


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listmgr.nrao.edu/pipermail/fitsbits/attachments/20250225/1935ed99/attachment.html>