[fitsbits] FITS conventions update

[Maren Purves]

On 12/15/14 07:36, William Pence wrote:
> On 12/15/2014 6:04 AM, Lucio Chiappetti wrote:
>> On Tue, 9 Dec 2014, William Pence wrote:
>>> This is to announce that a new FITS 'Header Space' convention has
>>> been submitted to the Registry of FITS Conventions,
>> I have [a comment] concerning the last sentence in the submitted text,
>> i.e.
>>
>> "It should be noted that if a FITS file is processed by software that
>> does not support this convention, then new keywords may be written at
>> the location of the END keyword (i.e., after the blank keyword
>> records). This will make the blank keywords unavailable for future use
>> by software that does support this convention"
>>
>> I wonder whether one might not do an update to the Standard, which
>> says that FITS *writers* are allowed to write new keywords over any
>> completely blank card image wherever it is located in the header (they
>> might even be allowed to "delete" a keyword overwriting it with an
>> entirely blank card image).
> I think that doing this would have undesirable consequences.  Many FITS
> files contain blank card images throughout the header that are used as a
> sort of visual separator between logically related sets of keywords.
> These blank keywords can be helpful for human readers of a header
> listing.  These embedded blank header records should probably  be
> considered significant, since they were placed there intentionally.  For
> this reason, this convention only treats the blank header records
> immediately preceding the END keyword as non-significant scratch space
> that can be reused when writing new keywords.

This is the case e.g. at UKIRT. Our observers and staff astronomers
want to be able to look at the headers and find related items grouped
together and separated by a blank line and a comment line that says
what the next section contains (e.g. weather parameters). Us software
people didn't think of that at the time (despite having astronomy
backgrounds) but the non-software astronomers told us otherwise.
People look at listings of FITS headers, not only software.

Thanks,
Maren Purves