[fitsbits] further reopening of Public Comment Period on the CONTINUE convention

[Tim Pearson]

The maximum number of characters in a string-value is an issue of semantics, not syntax, and depends on the usage of the keyword. With the current FITS standard, if no maximum length is specified in the appropriate document that defines the keyword, it is safe to assume that the maximum is 68 characters. If CONTINUE is added, there is no explicit maximum.

I therefore recommend that if this change is adopted, the rest of the standard be modified to specify the maximum length allowed for the string value for each keyword for which the usage is defined in the standard. The standard should also recommend that other documents defining string-valued keywords should specify the maximum allowed length. If this is a number > 68 (or explicitly "unlimited") then programmers will know that they have to be prepared to use the CONTINUE convention. If the appropriate document does not define a maximum length, then the backwards-compatibility requirement in FITS means that the maximum length is 68.

I disagree with the injunction "Be conservative in what you send, be liberal in what you accept". Being "liberal in what you accept" means that problems may go undetected for a long time before a less-liberal program trips over them. (I've encountered this many times: the offenders say "but DS9 accepted my file!")

Tim Pearson



> On Apr 21, 2016, at 11:43 AM, William Pence <William.Pence at nasa.gov> wrote:
> 
> I'd like to take a step back to view the bigger picture here:
> 
> Simply put, the proposal that we are discussing here is to incorporate the CONTINUE convention into the FITS standard in order to provide a way to record FITS string-valued keywords that are too long to fit on a single header record.  The CONTINUE convention has been successfully used for more than 22 years by several major observatories and it is supported by many of the commonly used FITS software libraries.
> 
> The motivation for developing the CONTINUE convention in the first place was to provide a standard way for projects to read and write long-string valued keywords in support of their own individual project or mission. The CONTINUE convention was never envisioned to be used with any of the approximately 30 string-valued keywords that are defined in the FITS standard.  Indeed, as far as I am aware, most FITS users find the current 68-character limit more than adequate for all of these keywords.
> 
> Since allowing the CONTINUE convention to be used with the mandatory and reserved keywords in the FITS standard would provide little or no added benefit and would potentially cause problems for existing software that reads those keywords, it was suggested to add the following sentence to end of the proposed text (which can be seen at http://sax.iasf-milano.inaf.it/~lucio/FITS/Conventions/continue-rev2.pdf):
> 
> "The CONTINUE keyword must not be used with of any of the mandatory or
> reserved keywords defined in this standard unless explicitly stated
> otherwise."
> 
> Assuming that this sentence is added to the proposal, does anyone have any remaining issues with the proposal as it is currently written?
> 
> -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)
> 
> 
> _______________________________________________
> fitsbits mailing list
> fitsbits at listmgr.nrao.edu
> https://listmgr.nrao.edu/mailman/listinfo/fitsbits