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

[van Nieuwenhoven, Richard]

I think this is a very good point! I could not write till now due to
personal problems, but this is also a deep conflict for fits-libs.

For example nom-tam-fits (and probably many other libraries) uses
long-strings dependening on the fact if the string (together with the
comment) fits as a normal string or needs a long string to store the data.

Now what should the lib do in future?

1. Throw an error if long-strings are used in a special set of keywords?
That would probably make a lot of fits files illegal that where legal
before (assuming longstring support was aktive).

2. Add some special api to deacivate the use of long-string in these
cases? That would mean changing the behavior of an existing api and make
the programs using it, not fits compatible.

Even when using a lib that currently forces the user to select between a
long-string or a normal string, the program itself will have some logic
that selects when to use long-strings. And this logic is different in
every single program....

If fits really wants to restrict the use of CONTINUE, than it should
exactly specify the keywords where not to use it. I do not think there
are a lot of keywords where CONTINUE would cause problems.

In any case, the CONTINUE feature will break some
fits-files/fits-programs in one way or another (in cases where the
String has a semantic meaning, or the string is copied to somewhere
else). I think we can only try to minimize the breakage by carefully
select a minimal set of keywords that would case "big" problems.

-Ritchie


Am 2016-04-19 um 06:19 schrieb William Pence:
> The IAUFWG suspended its vote on the CONTINUE convention a couple weeks
> ago to allow for further public discussion here on FITSBITS about some
> remaining issues.  It seems odd, therefore, that no significant
> discussion has taken place.  If anyone has any remaining objections
> about the wording of the current proposal (which can be viewed at
> http://sax.iasf-milano.inaf.it/~lucio/FITS/Conventions/continue-rev2.pdf),
> then please state them is here on FITSBITS, so that the wider FITS
> community can participate in the discussion and contribute to finding a
> satisfactory resolution to the issue.
> 
> At least one of the issues seemed to concern the possible negative
> impact on existing software that could occur if some of the currently
> defined string keywords (e.g., OBJECT or TELESCOPE) were allowed to be
> continued to have a length greater than the current maximum length of 68
> characters.  To mitigate this potential problem, I recently suggested
> adding a restriction on the use of the CONTINUE convention, such as:
> 
> "The CONTINUE keyword must not be used with of any of the mandatory or
> reserved keywords defined in this standard unless explicitly stated
> otherwise."
> 
> Would adding this restrictive language to the current proposal make it
> acceptable to everyone?  If not, please speak up.
> 
> -Bill