New UCD1+ for SIA protocol

[Robert Hanisch]

I just want to add my support for Alberto's objection to instr.scale -- I
think this could be very confusing.  The scale here is not an attribute of
the instrument, but of the image.  Glancing quickly through the UCD-UCD1+
mapping, however, I do not see a UCD that fits very well.

Is meta.bullshit going to stay?

Bob

----- Original Message ----- 
From: "Doug Tody" <dtody at aoc.nrao.edu>
To: "Alberto Micol" <Alberto.Micol at eso.org>
Cc: "Roy Williams" <roy at cacr.caltech.edu>; <dal at ivoa.net>; <ucd at ivoa.net>
Sent: Wednesday, March 24, 2004 4:52 PM
Subject: Re: New UCD1+ for SIA protocol


> Hi Alberto -
>
> On Wed, 24 Mar 2004, Alberto Micol wrote:
> > Only two points (in the temporary absence of an established data model):
> >
> > > VOX:Image_Scale
> > > --> instr.scale
> > > with the array value giving the scale in degrees per pixel of each
> > > image
> > > axis.
> > >
> >
> > Wouldn't it be better to have it as --> pos.wcs.scale ?
> > This parameter has to do with the way the data are sampled,
> > and any software could resample it (eg drizzle) onto a different grid
> > than the original instrumental pixel scale (and think of the case where
> > data from different instruments are mosaic'd together).
> > In the end the instrument (instr.) has nothing to do with the pixel
> > scale.
> >
> > And, anyway the pixel scale can be computed from the pos.wcs.cdmatrix...
> > why do we need it? Probably for queries ...
>
> All we are doing here is normalizing the UCDs for SIA - for V1.1 we would
> prefer to minimize serious changes to the interface.  Anyway, although
> the image scale can be computed from the CD matrix, 1) scale is a major
> image attribute which we prefer to have readily available to the client
> without having to process the WCS, 2) while the image scale is a required
> parameter, the WCS is merely "strongly recommended" and data providers are
> not required to provide it.  If the WCS is missing, it can be approximated
> from the required image parameters, i.e., CRPIX the image center, CRVAL
> is POS in IRCS coordinates, the image is assumed to be nonrotated, and
> the CD matrix is thus defined by the naxis and scale values.
>
> > Actually, where is the resolution in this scheme?  > Probably it was
> not in the SIAP, but I should read the SIAP specs again.
>
> SIA does not currently have the full sampling/bandpass data
characterization
> which we have discussed elsewhere.  At some point we would like to add
this.
> V1.1 is conceivable if it comes along fast enough.
>
> > > VOX:Image_AccessReference
> > > --> meta.ref.url
> > > specifying the URL to be used to access or retrieve the image.
> >
> > This is complex ... there could be various URLs associated to
> > the same image, e.g., to access the full product or only a preview;
> > but I need to read the SIAP doc again before insisting on this.
>
> These each get a different row of the query response VOTable, hence each
> has a different acref URL.  We have another proposal (again from Roy)
> to define an optional logical name which can be used to tag each row in
> the query result.  Rows which refer to the "same image" but differ only
> in image format, bandpass, or whatever, would have the same logical name
> to indicate that they are associated.
>
> - Doug
>
>