[fitsbits] FITS MIME RFC ready for review

[Tim Pearson]

On Tue, 3 Jun 2003, Steve Allen wrote:

> In a private e-mail exchange that happened before this posting was
> approved for sci.astro.fits it has become clear that the issue could
> be resolved by a one word change to the Internet Draft.
>
> Section 5.2 is the IANA application for media type "image/fits".
> The "Additional Information" starts with this sentence:
>
>     A FITS file described with the media type "image/fits" MUST have a
>     primary header and data unit (PHDU) which consists of at least one
>     pixel.
>
> The change would be from "MUST" to "SHOULD".

This change would mean that *any* valid FITS file could be described
as either image/fits or application/fits at the user's discretion.  Is
this the intention?

If any FITS file could have either media type, agents must be prepared
to treat the two media types as synonymous. So why do we need two
media types?

I thought the intention of image/fits was to indicate that the file
contained something that could be displayed as an image. A FITS file
with any of its NAXISn keywords zero does not look like an image to
me. If NAXIS1=0, it is of course a random-groups file.

Tim Pearson
Astronomy Dept 105-24, Caltech, Pasadena, California 91125, USA