Registry Use Cases

• These are use cases for data discovery
• Many of these are all-sky or large-area problems so that even after resources have been identified using the Registry, selective searches rather than cone searches would be preferable, see ADQL Use Cases MultiConditionCatalogueQueries (more to come).
• Cases where all requirements are already covered higher up are omitted. Hence for any given case all the requirements are not listed.

Requirements extracted from science cases

1. TopologiesImperialGroupA
   • Want catalogues containing redshifts
     • Need to be able to search specifically column headings or better still UCDs, not get back tables where redshift is in general description only. Keywords might also work.
2. GouldBelt and StellarProjectImperialGroupB
   • Want IR data in several large (tens-hundreds deg^2) regions
     • The Registry standards provide for categories Radio, sub-mm, IR etc. and also for wavelength limits. Would a tool to derive what the Registry wants from common, easily translated (at that sort of level) info e.g. frequency limits, band name, into what the Registry needs, help?
     • How to describe collections like INT-WFS with varying spectral and spatial coverage: which regions have been observed in which bands (i.e. within which wavelength limits); at present are spatial, spectral and temporal coverage all independent in Registry?
   • Coverage (and searches) may be more convenient in e.g. glat and glon
   • IRAS in B1950
     • Both points above - need coord conversion tool for Registry population and searches?
3. FIREllipticalsGroupC
   • Coverage of survey images
     • Region of Regard?
   • Is data science-ready, i.e. calibrated
     • What is 'calibrated'? Free from artefacts? In comparable units? (in both cases, to some degree, i.e. flat-fielded or etc, not necessarily perfect) - basically, can you work on it without instrument-specific tools?
   • Spectra needed * Does Registry always record type of data, incl. if spectra can be extracted from other kinds of data.
4. DeepFieldSurveys
   • The main data sets are well-known but some partial surveys are also useful (e.g. redshifts); on the other hand most data returned e.g. from an AstroScope query on HDFN is far too shallow or low-resolution data which just happens to cover the region.
   • Need to know resolution and sensitivity
     • Already provided for in standards
5. GalaxyClustering
   • Photometric redshift determination
     • Need to find data with accurate optical and possibly IR photometry
6. SupernovaGalaxyEnvironment
   • Need resolved images and/or catalogues with morphological information
   • Selection of astronomical data by time (e.g. before 2000).
7. BrownDwarfSelection
   • As well as nominal sensitivity, would be useful to know what is faintest object actually detected (and conversely the brightest, for the hot star case).
8. GouldBelt
   • Astrometric accuracy
   • Catalogue measurement aperture/image resolution

Other Registry issues

• Error responses to astronomer and to engineer (probably different)
  • Fault in query
    • Syntax etc.
    • Invalid (e.g. -ive RA)
  • Fault in AstroGrid service used for registry search
  • AstroGrid service inaccessible
  • Fault in endpoint of query e.g. AstroGrid Registry
  • Endpoint inaccessible
  • Information returned from endpoint is incomprehensible

-- AnitaRichards - 26 May 2006