Документ взят из кэша поисковой машины. Адрес
оригинального документа
: http://www.adass.org/adass/proceedings/adass03/P4-28/
Дата изменения: Sat Aug 14 04:02:50 2004
Дата индексирования: Tue Oct 2 05:35:54 2012
Кодировка:
Поисковые слова: п п р р р р р р р р р р р р р п
|
Next: The NOAO Mosaic Pipeline Architecture
Up: High Performance Computing
Previous: The AIPS++ Project
Table of Contents -
Subject Index -
Author Index -
Search -
PS reprint -
PDF reprint
Lemson, G., Dowler, P., & Banday, A. J. 2003, in ASP Conf. Ser., Vol. 314 Astronomical Data
Analysis Software and Systems XIII, eds. F. Ochsenbein, M. Allen, & D. Egret (San Francisco: ASP), 472
A Unified Domain Model for Astronomy
Gerard Lemson
Max-Planck-Institut für extraterrestrische Physik, Garching, Germany
Patrick Dowler
National Research Council Canada, Victoria, BC, Canada
A. J. Banday
Max-Planck-Institut für Astrophysik, Garching, Germany
Abstract:
We propose a framework for constructing a unified, conceptual
domain model for astronomy. We believe such a model to be an
essential ingredient for a future Virtual Observatory (VO). We also
give a high level, skeleton proposal for this VO domain model. We
indicate where details must be filled in for more specialized
models. We describe one detailed base level model, a component
model for Quantity, which is a generalisation of previous informal
proposals. Our domain model puts it in a larger context that
includes such concepts as measurement, error, and units.
Standard methodologies (see Fowler 1997, Booch 1994, Halpin 2001)
identify various phases in the software development process. The
first phase analyses the universe of discourse (UoD), that
is the world that we are interested in talking about in the
context of a particular project (Halpin 2001). The goal is to come
to a comprehensive domain model containing all the relevant
concepts and their interrelations. There are reasons why such a
model is an essential ingredient for the development of a Virtual
Observatory (VO). It will provide a common grammar and vocabulary
for expressing the varied data products existing in distributed
astronomical databases. It will define the set of concepts that a
user of the VO can use in queries to these archives. Without such
a common, ``Esperanto'' data model it will not be possible to
achieve true interoperability between different archives. In
different terms, we believe the conceptual model arising from the
standard analysis phase to be equivalent to an ontology.
We define the UoD for the VO as ``the work that astronomers,
astrophysicists and support scientists do and the results they
have obtained''. Our motivation for this choice is that we believe
that users of the VO are ultimately interested in the results of
the work done by other astronomers. Users are not
``just'' interested in getting access to images, simulation results
or other physical results of astronomical research, stored in some
astronomical archive, but will want to know what is
actually represented by these results, how they were
obtained, what experiments were executed and how. The latter is
what we mean by the term ``work''.
When we say that we believe VO users will be interested in the
experiments that produced the results, we mean that they
should be interested in them. One of the main
tasks of the VO is to enable other astronomers to do rigorous
science with the results and services that are made available through
it by their colleagues. It is obvious that results can
only be interpreted through knowledge of the process that produced
the results - the ``provenance''. We believe the VO has both the chance
and duty to formalize the concepts underlying this
provenance by including them explicitly in the modelling effort.
Here we list some of the core concepts we believe need to be
modelled explicitly to give a proper description of the world of
astronomy, the UoD of the VO. In the full model
(http://www.g-vo.org/materials/UDM-Poster.pdf), these
have been translated into UML classes and worked out in more
detail.
- experiment
By experiment, we mean that work that
leads to results that are of interest to users of the VO.
- result
The data resulting from astronomical work that users of the VO are
assumed to be interested in.
- protocol
It should be possible to retrieve what was actually done,
why and how. This we call the protocol of the experiment.
Separating this concept from experiment has an
obvious use when a certain protocol is followed in multiple experiments,
eg. using an analysis tool such as SExtractor.
- objective/observable
The goal of the experiment, as described in the protocol. The objective
of an experiment is to assign values to variables.
- variable
One can assign values to variables by measuring them or
calculating them. Thus, the variables are the things which are to be or which
have been determined by an experiment.
- measurement
A result consists of values that have been assigned to variables.
A measurement is a special kind of ``value assignment", namely the
common one corresponding to a numerical value.
- value (quantity, category)
This is the smallest atomic unit in the result hierarchy. The
concept value corresponds to a value assigned to a property. SI
calls this a ``value of a physical quantity'' (NIST 1999).
- uncertainty/confidence/error
In the scientific process of assigning values to properties
by measurements, errors will be made. These may be
statistical, systematic, and possibly correlated. However they are caused,
to interpret measurement values it is imperative that they be
accompanied by the confidence/uncertainty associated with them.
- phenomenon
This concept corresponds to SI's ``quantity in the general sense'',
or to Fowler's PhenomenonType (Fowler 1997). It is the
generalization of ``a property that a body or substance can have"
(NIST 1999). Phenomena are the ultimate things scientists are
trying to determine.
- property
This is a phenomenon assigned to a particular thing, as in
``a property of'' that thing. For example, colour is a phenomenon
and the colour of that galaxy is a property of that galaxy.
- subject
Subject is the concept that unifies the thing
mentioned above and in the SI definition for ``quantity in the
general sense". It includes other concepts like body, substance,
region, etc. It is really an anchor concept that corresponds
to a collection of properties, since a particular thing
is defined by properties that we can measure, know, or be interested in.
- unit
Unit defines the ``measuring rod'' that is required to interpret
numerical values (quantities) assigned to numeric properties. SI
defines it as a particular property of a particular, identified
subject. All values of the same phenomenon can be expressed by
giving the ratio of that property to the property of the
identified subject.
- reference system
Just as a unit specifies the measuring rod for interpreting
numerical values assigned to a property, so the reference system
can be thought to specify the zero point. Examples include
a coordinate system for positions on the sky, and a magnitude
system for expressing fluxes.
- standard
To enable interoperability, or more simply, for us to be able to
understand each other's (meta-)data, it is not sufficient to
provide a model (the grammar of the ``Esperanto" discussed above),
we also need to agree on a number of standard instances - the
``vocabulary".
- physical artifacts (database, file)
Data will be stored in some physical datastore which may be a
filesystem or a more formal database system. We need to be able to
identify and locate these containers.
Figure 1:
A detail of the
domain model, dealing with measurements and values.
![\begin{figure}
\plotone{P4-28_1.eps}
\end{figure}](img1.gif) |
Modelling quantitative values with units and errors has received
considerable attention in the IVOA data modelling working group. In
Figure 1 we propose a model for this area.
The main difference between this and other proposals
(see http://www.ivoa.net/twiki/bin/view/IVOA/IVOADMQuantityWP
for IVOA Quantity data modeling resources )
is that we believe
that the core concept is the measurement, which we define as
the act of assigning a value plus error to a property, not the quantity
itself. We further generalize the Quantity concept to values and classifiers.
To see how this model fits
within the large scale framework we refer the reader to the full
diagram (http://www.g-vo.org/materials/UDM-Poster.pdf).
Acknowledgments
This work was undertaken as part of a Canadian Virtual Observatory
(http://services.cadc-ccda.hia-iha.nrc-cnrc.gc.ca/cvo/,
CVO)
and German Astrophysical Virtual Observatory
(http://www.g-vo.org, GAVO)
collaborative research project. CVO is sponsored by the National Research
Council (NRC) and the Canadian Space Agency (CSA). GAVO is sponsored by the
German Federal Ministry for Education and Research (BMBF).
References
Grady Booch 1994,
Object-oriented Analysis and Design, 2nd edition, Addison-Wesley
Martin Fowler 1997,
Analysis Patterns, Addison-Wesley
NIST 1999,
SI Specification,
http://physics.nist.gov/cuu/Units/introduction.html
Terry Halpin 2001,
Information Modelling and Relational Databases, Morgan Kaufmann
© Copyright 2004 Astronomical Society of the Pacific, 390 Ashton Avenue, San Francisco, California 94112, USA
Next: The NOAO Mosaic Pipeline Architecture
Up: High Performance Computing
Previous: The AIPS++ Project
Table of Contents -
Subject Index -
Author Index -
Search -
PS reprint -
PDF reprint