[fitsbits] start of Public Comment Period on the INHERIT convention

[Phil Hodge]

On 06/23/2015 04:05 PM, Landsman, Wayne B. (GSFC-667.0)[ADNET SYSTEMS 
INC] wrote:
> In addition, eliminating the INHERIT convention would eliminate any 
> distinction between keywords with "fixed" image values, and those 
> whose values can change with each extension.     This meta information 
> in the INHERIT convention provides additional data integrity -- any 
> updates of the fixed keywords are forced to apply to every image 
> extension.
>

The proposal explicitly addresses the case of a keyword being present in 
both the primary header and in an extension header, and even if INHERIT 
= T, the value in the extension header takes precedence.  That is, the 
proposal for INHERIT does not force the primary header value to be 
adopted for every extension.

> I am not certain if the advantages of INHERIT convention are 
> sufficient to include it in the FITS standard.      But I hope that 
> groups ( such as STScI  ) that currently use the convention will 
> continue to use it .      --Wayne

I hope not.  I was one of the people at STScI who (20 years ago) thought 
this was a good idea, and I apologize.  If you want this feature, it's 
easy:  Look for a keyword in an extension header; if the keyword is not 
present, look for it in the primary header.

Phil