What do the 1990s hold for astronomy
and astrophysics?
What do the 1990s hold for astronomy
and astrophysics?
Editor's Note: Every 10 years the
National Academy of Sciences and
the National Research Council charter
a committee to survey and
prioritize
the most important programs in astronomy
and astrophysics
for the next decade.
The committee to study the 1990s
was chaired by
John N. Bahcall of
Princeton's Institute for Advanced
Study, and its
report was entitled
The Decade of Discovery. NCSA Director
Larry
Smarr chaired the 26-member
Panel on Astronomical Computing and
co-authored a journal article with
William Press, professor of astronomy
and physics at Harvard, on the panel
report of the Bahcall Committee
(Computers
in Physics, March-April 1991). The
following questions, compiled by
Fran Bond, access editor, and answered
by Smarr and NCSA
Research Associate
Radha Nandkumar, summarize the panel
report and
are based on the Computers
in Physics article.
1. What major trends do you see developing
in the 1990s for astronomy and astrophysics?
The High Performance Computing and
Communications Program, the
national
multiagency cross-disciplinary initiative,
will pose a new context in which
scientific research will be carried
out. Astronomical science--by virtue
of its intrinsic data and computation-intensive
nature, its manageable size as a
discipline, and its past
experience
and future opportunity--stands poised
to be the cutting-edge application
discipline in a number of major aspects
of a
national program.
In three complementary areas (digital
data acquisition and archiving,
intensive data processing, and theoretical
modeling), astronomers and astrophysicists
are ready to take advantage of virtually
all of the expected technological
advances of the 1990's: desktop
high-performance workstations, widespread
use of parallel computers, large
increases in memory capacity, improvements
in data storage technologies, improved
use of graphics and visualization
techniques, high-speed networking,
and powerful new algorithms. The
principal goal for astronomical computation
in the1990s is connectivity of all
these elements to guarantee access
to the
appropriate level of computational
resources for all researchers.
2. During the 1980s the NSF supercomputer
centers made
supercomputers available
to the research community and a significant
segment took advantage of the opportunity.
What will be the centers' role in
the 1990s? And what did the committee
recommend?
Despite the advances in desktop workstations
and local computing
resources, there
will still be a place for large central
computers:
super-computers will
be bigger and faster versions of
what sits on a
desktop. These machines
with their huge memories, extensive
disk
capacity, and extremely fast
processing and input and output rates,
will be crucial to a subset of both
the theoretical and observational
computational users with the most
demanding problems. These are the
Grand Challenge problems in astronomy
and astrophysics. The committee recommended
that a hierarchy of computing resources
be made available to the general
astronomical community under the
assumption that the supercomputer
centers, which must be periodically
upgraded to remain at the forefront
of technology, continue to provide
this scarce resource to the community.
3. Will desktop computing continue
to be as significant in the nineties
as it was in the eighties? If so,
what changes do you foresee in the
role of the desktop computer?
The 1990s will bring major advances
in all levels of the computing hierarchy.
The desktop environment and high-speed
networks will alter the mode and
methodology of the way research is
accomplished. New
microprocessor
technologies are already positioning
affordable, powerful computers on
every astronomer's desk. These will
be linked
to each other over a national
network, as well as to high-value
resources such as supercomputers,
national observatories, and national
data banks. Some of these machines
are nearly as powerful as today's
supercomputers for many tasks. These
workstations will make possible the
acquisition and processing of large
datasets, will provide links with
research bibliographies, and will
allow for the comparison of theoretical
models with observations.
4. With the advancement in computational
astrophysics/astronomy in the past
10 years, are there any long-standing
problems that are nearer solution
in the coming decade? If so, what
are some of them?
The 1990s will be the decade where
a number of long-standing astrophysical
problems will be solved and computers
will play an
important role in these
solutions. Areas which seem particularly
ripe
for rapid theoretical progress,
and comparison with observations,
can loosely be categorized as follows:
large-scale structure of the
universe
and cosmology, active galaxies and
jets, star formation and
the interstellar
medium, dynamics of stars and stellar
atmospheres, supernovae, accretion
onto compact objects, generation
of
gravitational radiation, and
the microphysics and magnetohydrodynamics
of astrophysical plasmas, and the
addition of
nonequilibrium physics
to hydrodynamics codes.
5. What communitywide recommendations
were made in the area
of software
and code development?
In order to utilize effectively the
enormous advances in computer
hardware
and in digital detectors expected
in the next decade, there must be
an accompanying development of scientific
software. The panel commended the
observational community's leadership
during the past decade in the development,
maintenance, distribution, and support
of community software such as AIPS,
IRAF and FITS. NSF is commended for
the support of this effort. The
panel recommended that innovation
must be fostered in the observational
software arena as new detectors and
new computer architectures create
new opportunities for community tools,
and that the scope should be broadened
to include the development of community
software for theoretical modeling.
6. With the number of space flights
launched by NASA, there must be a
tremendous data archive now. And
widening the capabilities for archiving
data is of general interest
today.
What did the committee advise in
this area? Would any advancements
in data archiving spinoff to the
national
centers? And would researchers
at smaller institutions
share in
this?
Current and forthcoming missions
in NASA's space astrophysics program
may result in an average archival
data rate of close to 10 terabytes
per year by the end of the decade.
Once the data have been processed,
the
volume that is distributed in the
community will be a multiple of
the
primary archivally stored data. An
appropriate long-term goal is for
all electronic data obtained from
space-based and ground-based telescopes
to be archived and catalogued using
modern data base management systems
and for the scientific community
to have free access to the archives.
A nationally supported archiving
program will
allow researchers,
students, and smaller colleges and
universities to enter the mainstream
of modern research in a way that
is not now possible. Research using
astronomical data archives can thus
be very broadly based.
7. Could you summarize the panel's
recommendations in a
brief overview?
Of these recommendations, which ones
deserve top priority? The Astronomy
and Astrophysics Survey Committee
made major
recommendations based
on input from the panel on Computing
and Data
Processing, which made
specific recommendations in four
areas: archiving, workstations and
hierarchical computing, networks,
and
community code development.
The committee recommended:
- Digital data archives encompassing
ground-based and space-based
observations
should be established and made available
over high-
speed national networks.
- Individual workstations and departmental
minisupercomputers should be acquired,
providing a distributed network for
astronomical computing.
- NSF should encourage the development
of high-speed national networks
by funding links to widely used observatories.
- NASA should increase its role in
furthering the development of community
software and standards.
Regarding archiving, the panel made
additional recommendations, including:
- All major ground-based observatories,
both public and private, should
incorporate the capability for archiving
of digital data.
- Archives should be designed to
outlive any specific hardware, software,
or media.
- Archives should include all the
raw data, calibration data, and information
necessary to remove instrumental
signatures from the data.
- Policy should be established to
require the archiving of data described
in papers published in the refereed
literature. This
policy should be
enforced and implemented through
journals, time-allocation committees,
and the proposal reviewing process.
Regarding community code development,
the panel's recommendations were
even more specific than the committee's:
- NASA and the NSF should support
the development of community codes
by funding proposals selected by
peer review in an AO/NRA process.
NASA should expand the initial effort
of the Astrophysics Software and
Research Aids NRA. There should also
be support for the maintenance of
codes found to be of general utility
to the community.
- A community code program should
fund both individuals who propose
small programs and larger groups,
consisting of both computational
astronomers and software professionals
of high quality, who propose major
efforts.
- A recommended level of effort could
involve the funding of three large
groups (each with five to eight professionals)
and a number
of smaller groups.
__________________________________________________________________
access / May-June 1992 / NCSA / pubs@ncsa.uiuc.edu