Registry Curation Challenges

1
THE US NATIONAL VIRTUAL OBSERVATORY
Registry Curation Challenges Bob Hanisch US NVO
Project
2
Why Do We Need Curation?

• Resource metadata are not equally well understood
  by data and service providers
• Registry interfaces have varying levels of
  sophistication, providing more or less guidance
  in creating metadata
• Some metadata is not immediately
  availablerequires astronomical experience to
  characterize resource properly
• Automated metadata population incomplete
• Poor quality of metadata in registry will
  undermine its value

3
Curation Examples

• NVO Registry Query http//nvo.stsci.edu/voregistr
  y/index.aspx
• CHANMAST
• UIT (MAST)
• Hubble UDF
• Vizier tables

4
CHANMAST from HEASARC
Confusing name
Not a well-formed identifier
Ambiguous
No instrument name
???
No facility name
5
UIT from MAST
Vague
Wrong
Unspecified
Wrong interpretation of Facility
6
UDF from STScI
Poor choices
No
Not useful
7
Vizier tables from CDS
Mix of Title and Creator content
Useful?
Not proper identifier
Buried in Title
Better label if simulation?
Missing
8
Registry Maintenance

• Harvesting
• Frequency
• Reliability
• Curation control
• Replication and redirection
• Authority