UCDs

1
UCDs

• Notes on discussion
• 2003 October 16
• 1400 1600

2
Roy Williams and Sébastien Derriere

• Introduction
• Need for revision of current UCD scheme
• Presented UCD working draft V1.9.9
• General form
• ltnamespacegtltconceptgtltpropertiesgt
• Examples
• arith.radiophot.fluxem.radio
• Ivoameta.idsrc

3
Wide-ranging discussion

• How to support covariance, e.g. proper motion and
  declination?
• Where is boundary between UCDs and Data Models?
• Ideally need normalised vocabulary, but IAU
  Thesaurus is 10 years out-of-date.
• What happens to UCD if columns are combined,
  e.g. arith.sum ?

4
Patricio Ortiz

• Presented analysis of most popular UCDs showing
  the complexity of the problem.
• This was followed by a further inconclusive
  discussion. Some topics raised
• How/whether to support grouping generally felt
  to be responsibility of data model UCD should be
  applied to a column as an entity.
• Is equinox needed as attributed to positional
  coordinates?
• Can UCDs be applied to parameters as well as
  columns?

5
Tom McGlynn

• Proposed somewhat simplified scheme for UCD2
• Fewer top-level items, mainly physical concepts
  such as position or flux.
• To be followed by one or more attributes/qualifier
  s
• Waveband, measurement type, etc.
• Value (this one alone may be omitted as the
  default)
• Error, max, min, etc.

6
Straw vote taken

• Do we need new UCDs, or keep old ones
• In favour of existing UCDs 4
• In favour of some form of new UCDs 25
• There were also a few abstentions.

7
Demonstration of Aladin by Pierre Fernique

• Showed Aladin using UCDs to limit the number of
  stars plotted on an image
• Select PHOT lt 16
• Can also be done in UCD-2, perhaps even more
  effectively.

8
Conclusions

• A move to some form of UCD-2 was generally
  supported.
• The simplification and rationalisation proposed
  by Tom McGlynn attracted relatively few
  criticisms and several favourable comments.
• Coordination is desirable on the appendix to
  UCD-1.9.9 which has a table of waveband names,
  e.g. em.UV.200-300nm
• These should be aligned with the waveband names
  or classifications proposed by the Registry and
  Data Model groups.