more on FITS and VO
fitz at noao.edu
Fri Nov 28 13:11:10 PST 2003
I think we're saying the same thing, more or less: extensions
belong in application/fits services and primary arrays in image/fits.
I'm in favor of the wording change since requiring the PHDU pixel forbids
zero NAXIS(n) values if you want to send a compliant FITS file and so
you can't send a dataless header. Allowing extensions in image/fits
would allow me to send a mosaic MEF to be displayed by DS9 or somesuch
which understands the file, but also opens a can of worms about what
current and future extensions could also be put in the file which may
break stupid viewers. To me, image/fits is a simple 2-D array that can
be displayed, application/fits is the possibly more complicated format
destined for a task/package that can do smart things like ignore extensions
or operate on the data as a whole.
More information about the interop