🚀 go-pugleaf

In article <1cibc0INNpjq@darkstar.UCSC.EDU>, sla@umbra.ucsc.edu (Steve Allen) writes...
>I am looking through the NOST 100-0.3b standard document trying to be
>very sure about one point:
> 
>    Section 5.3.2.1 specifies that a "character string shall be composed
>    only of ASCII text."
> 
>    Section 6.1 specifies that characters shall be 7-bit ASCII
> 
>These specifications do not preclude the existence of unprintable characters
>(ASCII 0 through 31 and 127) in a character string.  Is it intended that
>unprintable characters be legal?  Are there FITS readers which will be
>confounded by unprintables?
> 

I've always gone by the philosophy that only ordinary printable characters
should appear in a FITS header.  For instance, my software will expand tab
characters out to the appropriate number of spaces before writing it into the
header.

One might argue that more leniency could be permitted in ASCII fields in FITS
binary tables.

>A further question might be asked about the reason why 8-bit ASCII is not
>allowed, but I would rather defer that until later.

That's simple:  8-bit ASCII isn't really a standard, is it?  You get different
characters on different machines.

Bill Thompson