Well yes, we know a degenerate axis is not the right way to do it, but
my point was that there are never-the-less lots of files out there that
use this convention.
David
On Wed, 6 Sep 2006, Malcolm J. Currie wrote:
> > Having a redundant STOKES axis seems pretty common.
>
> STOKES parameters shouldn't be along an axis. Before extensions were
> invented one could see why this convention was adopted. These days each
> array ought to be in an IMAGE MEF or altogether a binary table with
> appropriate per-array headers in the table too. Also many packages out
> there will flounder if they encounter a four-dimensional array (or
> three-dimensional if expecting an image).
>
> There's an easy workaround with NDFCOPY.
>
> Malcolm
>
|