VOTable 2005

1
VOTable 2005

• Discussion Summary
• and Roadmap

2
Agenda

• http//ivoa.net/twiki/bin/view/IVOA/InterOpMay2005
  VOTable
• Description of some applications dealing with
  VOTable interpretation/transformation
• Nilesh Urunkar About C Parser and CONVOT
• Mark Taylor About STIL/STILTS/TOPCAT
• Improvement of the VOTable schema
• Gerard Lemson About VOTable Schema

3
Agenda (continued)

• UCD1, UCD1 and utype in VOTable
• Sebastien Derriere About UCDs and utype
• Jonathan McDowell Comment on utype
• Coordinate system in VOTable
• Arnold Rots About simple STC, LinearSTC?
• Discussion on implementations
• Open Questions
• Other datatypes (dateTime, string, ...)
• Future of the VOTable group ?

4
VOTable interpretation/transformation

• Interpret all serializations
• streaming / non-streaming modes
• full validator (not schema only)
• graphical interfaces available
• ? VOTable mature widely used

5
Changes in VOTable schema

• Construct closer to Java binding, no global
  element
• Single inheritance mapping, only extension, no
  substitutionGroup
• ? Change in VOTable schema, with PARAM inherited
  from FIELD (v1.11) (need for a Proposed
  Recommendation / Recommendation path ?)

6
UCD1 / UCD1 / utype

• Some applications still use (old) UCD1 syntax.
• ? make clear from next version (1.12 or 1.2?)
  that UCD1 syntax only is valid
• utype represents a pointer to some parameter of a
  data model
• ? use XPath-like syntax (write a short formal
  document with DM?)

7
Various questions

• Mime type for VOTable application/x-votablexml
  (experimental)
• Concepts of primary/foreign keys in VOTable (or
  in extension schemas?)
• other datatypes (date/time, string)? not yet
  (FITS compatibility)

8
STC in VOTable

• In VOTable coordinate system currently defined by
  a dedicated ltCOOSYS system"ICRS"
  epoch"J1991.25"gt
• ?to be replaced by STC definitions in order to
  keep common model across VO components
• Some problems in STC import in VOTable context

9
The STC problems

• ltAstroCoordSystemgt fine but
• necessity of defining unused systems (TimeFrame,
  SpectralFrame, RedshiftFrame)
• could be referenced via a utype ?
• ltAstroCoordsgt problems
• Epoch need frequently to be defined
• Possible conflicting units with ltFIELDgt
  (interpret as default unit?)
• Actual values are deferred via ltValue2Refgt which
  has to reference a vector of 2 numbers (in
  VOTable it's necessary to reference the other way
  around)

10
Possible solutions(private discussion A. Rots)

• Define the position as a VOTable ltGROUP
  ref"ICRSastroCoordPosition"gt referencing the
  ltAstroCoordgt part
• what happens if one component of position is
  missing (e.g. Dec in meridian observations)
• Refine the utype syntax to reference individual
  components of the STC model
• Other solutions (include STC within a VOTable
  ltGROUPgt ?)

11
STC in VOTable Roadmap

• Proposal need to be discussed between VOTable and
  DM (toward a VOTable V1.2)
• ? possible solutions written down in June
• ? discussions in summer/september
• ? report in Madrid