Unified Content Descriptors
Semantic types for astronomy
Roy Williams,
California Institute of Technology
History
• Created at CDS Strasbourg
• (the first VO prototype)
• Harvested
• From 5000 tables, 20000 table columns
• To create ~450 UCD words
• Example
– pos.eq.ra means right ascension
– think of a UCD as a “semantic type”:
int n;
float x;
pos.eq.ra alpha;
UCD is a semantic type (not a name)
example: tablae of double stars
name alpha1 delta1 alpha2 delta2
ucd pos.eq.ra pos.eq.dec pos.eq.ra pos.eq.dec
Summary
• UCD is inherently fuzzy
– does not contain all metadata
• UCD is a description, not a unique name
• UCD already works
– easy to shoot at!
• UCD is mined from large amount of metadata
– phenomenological, not opinion
Natural Data
Language UCD Model
Richer, Fuzzier Thinner, Exact
Why is UCD useful
• Finding table attributes
• Aladin can draw based on UCD info
• Cone search demands UCD labelling of output
• Searching a registry that has tables
– Relevance of a table based on its UCDs
– Registry metadata for some data types defined with
UCDs
• VOResource includes VOTable, which has UCD
description of table attributes.
• Do this with a Matching Function
– Which UCD matches best the UCD I am looking for?
• Data exchange from legacy systems
UCD and utype
• UCD is a type,
struct Myposition {
pos.eq.ra alpha;
....} standard meaning
• utype is a data-model-part
struct VOStandardPosition {
float RA;
standard meaning
...}
General Syntax of IVOA UCD
• [namespace:] w.o.r.d; [namespace:] w.o.r.d; ....
• Character set is
– A-Z, a-z, 0-9, hyphen, + 3 special
– No white space
– Case insensitive
• Three special characters
. period for hierarchy
: colon for namespaces
; semicolon for separating words
UCD Tree
arith Quantities related to arithmetic and mathematics, including count, difference, ratio
em The electromagnetic spectrum
meta Quantities related to metadata, such as identifiers, flags, notes, URL, and
instr Quantities related to an instrument; typical sub-levels are telescope, observatory, etc.
obs Observation methods such as detector, filter, plate, spectrograph, exposure time, etc.
phot All photometric measurements, organized according to the wavelength; includes polarization.
phys Generic physical quantities, such as length, velocity, mass, and including atomic & molecular concepts and
properties, temperature, pressure, gravity, etc...
pos Position in the sky, reference frames; including equatorial, galactic etc coordinates; geocentric, heliocentric
etc; and precession and nutation.
spect Quantities related to spectroscopic measurements
src Properties of the observed source of radiation: source classifications and morphology, extension in the sky,
variability,
stat Statistical quantities and quantities related to model fitting, including concepts such as error, maximum,
residuals.
time Quantities related to time.
Multiple words
First word: what semantic type is this thing?
-- other words are helpers
Quantity UCD
Magnitude phot.mag
Flag (weather, reliability, meta.code; phot.mag
etc)
Error on magnitude stat.error; phot.mag
Maximum temperature of an phys.temperature; instr;
instrument stat.max
Error of a V-band magnitude stat.error; phot.mag; em.opt.V
Names need a Namespaces
“We took the table and chair dimensions,
and wrote them in a table.”
Namespace =
furniture
Namespace =
wordProcessing
FITS keywords
have no What does
namespace! “TARGET”
mean?
Namespaces
• Namespaces
– Introduce special character colon :
– example: mynamespace:em.opt.myfilter
– Default namespace is ivoa
– Should be an attempt to integrate to core
– UCD committee
– Example: SIAP 1.0 has VOX namespace
– VOX:Image_AccessReference (should be meta.ref)
– VOX:WCS_CoordRefPixel (should be pos.wcs.crpix)
Cone Search
• service request is a cone
• service response is a table:
blahblah foo bar banana syzygy
meta.id pos.eq.ra pos.eq.dec
Must have 3 columns May be other columns
with these UCDs
UCD Use cases
• Subclassing a table
• Table that has columns of specified UCD
– eg Conesearch, SIAP
• Aladin
• [phot.mag*]
VOEvent
• Types of objects
– BH, burster, AGN, etc etc
• Types of events
– SNe, GRB, lensing event, flare, etc etc
• Use cases
– Shall I observe this/
– Probabilistic interpretation
– Has interpretation changed with follow-up?
Semantics
• RDF (Resource Description Framework)
• statements are triples:
– “Resource Related-to Resource”
– (Cat Is-instance-of Animal)
– (Sigma Is-standard-deviation X)
– (RA is-member-of position), (Dec is-member-of position)
– T has-UCD phys.temperature
• Could use UCD for deduction
– Create SED from magnitudes
– Grouping parameters to make complex object
» eg error ellipse as errX, errY, covXY
RDF
Resource Description Framework
is for describing and exchanging metadata
Example of an RDF statement
Roy Williams = Author ( http://www.cacr.caltech.edu/~roy)
object predicate subject
Resource Resource Resource
or literal
Roy Williams