HEASARC%20Experiences%20with%20UCDs - PowerPoint PPT Presentation
Title:
HEASARC%20Experiences%20with%20UCDs
Description:
Discovery: Need to find tables that address a particular question. ... E.g., for EGRET and GLAST, hard implies ~50 MeV not 500 keV. How do we do upper limits? ... – PowerPoint PPT presentation
Number of Views:21
Avg rating:3.0/5.0
Title: HEASARC%20Experiences%20with%20UCDs
1
INTERNATIONAL VIRTUAL OBSERVATORY ALLIANCE
HEASARC Experiences with UCDs
Tom McGlynn and Michael Preciado, NASA/GSFC
2
What do we see as the purpose of UCDs?
- Discovery Need to find tables that address a
particular question. - Analysis How do I understand a table that I have
retrieved. - Identification Mark specific columns in a table
(SIA, Cone search) - To be superseded by UTYPE?
- Concentrate efforts on columns of a table that
will be of interest generally.
3
Minor issues.
- Non-specific UCDs
- Meta.number, meta.note, meta.code
- Does this convey more information than just a
blank UCD? - Ambiguities
- When do we settle for an OK UCD, when do we need
to get a better one?
4
The Observing Process
- UCDs defined for object tables not for
observation tables. - Current UCDs handle the details of the actual
observation but not of the entire observing
process. - Proposed, scheduled, actual, ...
- Add adjectives like obs.proposed, obs.scheduled
(e.g., time.expoobs.proposed) with implicit
default of obs.actual or some such) - Times Readout time, download time, release
date,... - Probably requires new nouns
- Processing status, software configuration used,
calibration status, revision, ... - Does this belong in obs or perhaps a new
processing tree. - Obs.calib is already there but not clear how to
use it in this context) - Proposal cycles
- Information that links to archive
- Part of obs tree?
- Obs.data.filename (and maybe obs.data.filenameob
s.calib for calibration data?) - Datatypes, data sizes
- Maybe a generic concept here, modified by the
appropriate context - File.sizeobs.data.filename maybe
comp.file.sizeobs.data - These are concepts that users are very much
interested in...
5
Miscellaneous concerns
- Can we build new UCDs from existing words?
- Have UCDs for semimajor and minor axes. Do we
need to use different UCDs for major and minor
axes versus semi...? - Bands in X-ray and gamma-ray regimes. The
distinction between hard and soft gamm-rays does
not accord well with current usage. E.g., for
EGRET and GLAST, hard implies gt 50 MeV not 500
keV. - How do we do upper limits? How do we distinguish
upper limits from upper bounds? Stat.max for
both? - Meta.id versus meta.id.cross versus meta.id.assoc
how systematic before we use the second, how
distinct before we use the third? - Phys.absorption versus phot.color.excess
6
Misc. contd.
- Meta.bib versus meta.ref
- Why spect.dopplerVeloc.opt rather than
spect.dopplerVelocem.opt? - Counts/densities of nearby sources
- V/Vmax (some kind of stat but I dont know
what). - Opacity of source
- Instrumental characteristics versus instrument
settings - Can any numeric quantity actually be just
primary? - (Given X cant I always have stat.errX)
7
Summary
- Most columns we can find UCDs for.
- There are major holes in the handling of
observation tables - Hard to understand the real boundaries between
UCDs.
8
Concern
- How do I search for UCD matches.
- Can I do this mechanically?
- Does a.b match a.bc.d
- Does ac.d match a.bc.d
- Does making UCDs more specific make it harder to
get matches?
Recommended
CrystalGraphics Presentations
