NVO Metadata Working Group Meeting
2002-08-22  Thursday 11am ET (10am CT, 9am MT, 8am PT)
Call: 1-800-456-4561, or
      1-719-955-1115
Passcode: [contact Ray]

Attendance:  Gretchen Greene, Michael Louis, Jonathan McDowell, 
	     Ray Plante, Doug Tody, Roy Williams

1. Roll call & News
   None

2. Simple Image Access Document

   Pre-discussion:
   Roy:  His DSS service displays messages that indicate what's
	 happening as the image is retrieved.  Do we (fore)see a way
	 to expose this through the standard interface?  This seems
	 like the sort of thing we want to move to.

   Doug: A general purpose message is usually part of a distributed
         system, but we haven't really talked about it in NVO, yet.
	 It's probably beyond what we do in this initial spec.  It
         would be easier to address this with SOAP-based Web
         services. 

   Ray:  You can do some of this with the current spec.  You can
	 choose to implement retrieval in one of two ways.  First is
	 when the user requests a particular image, you can block
	 until its ready.  Second is that you can return a DEFERRED
	 status; this can include user-oriented statements (in the
	 content of the INFO element) about what's going on.  This is
	 latter scheme is a "pull" mechanism, while Roy's custom
	 service uses a "push" mechanism.

   Doug: What is the status of adding WCS info to DSS cutout headers?
	 This can be in addition to the plate solution info and is
	 straight-forward to convert.

   Roy:  pointer to software?

   Gretchen:  We've done this for our service.  We use software from
	      Mark Calabretta.  There is a pointer to it on our Web
	      page.  

   Ray:  Could you post that URL to the MWG list?

   Main discussion:
SIA document discussion:
- Doug sent out a status summary which he discussed.
- The issue of granularity:
 How do we handle bandwidths? Gretchen says they
  have services to do both 'all filters' and 'one filter'.
 Comment on Tom's Skyview and how it categorizes services
 according to waveband. It it easier to redo Skyview now?
 Suggestion to standardize waveband definitions to what
 Skyview does.
 Do we need waveband numerical value (center and width)
 as well as vague regimes?
-We'll use to specify boundaries btwn waveband regimes
 Propose to use Tom's bands, but argued that IR needs
 to be split up into FIR and NIR since the data is very
 different. 
 Consensus: keep names in current SIA spec, except replace "infrared"
      with fir and nir.  
- *Action JCM to come up with scheme for describing waveband info.
  - UCD names for columns provided, and type and units (Hz preferred?)
  - Include low, fiducial and high 
  *Action JCM and DT to look at WCS paper standard on spectra

- Discussed whether we need to specify a null value.  A
  self-describing mechanism for this is defined as part of the VOTable
  spec. 
- Commented that can add extra parameters to current spec
  to build images on the fly from e.g. X-ray data using
  an energy filter. On-the-fly images are regarded as
  examples of mosaicing services.

4. Review of Action Items

Gretchen: post URL for info on software for converting between DSS
          plate solutions and WCS.
Jonathan: propose columns for the VOTable response that describes the
	  frequency coverage.
Jonathan/Doug:  check FITS standard and proposals for standards for
	  expressing spectral information that we might want to be
	  consistant with.