perm filename CONFER[F89,JMC] blob
sn#878230 filedate 1989-10-10 generic text, type C, neo UTF8
COMMENT ā VALID 00002 PAGES
C REC PAGE DESCRIPTION
C00001 00001
C00002 00002 %confer[f89,jmc] Fears about the strategic directions conference
C00023 ENDMK
Cā;
%confer[f89,jmc] Fears about the strategic directions conference
Here is the message that has aroused my fear. It is followed by comments.
ā09-Oct-89 1853 @Neon.Stanford.EDU,@Polya.Stanford.EDU:THEORYNT@VM1.NoDak.EDU Strategic Directions in Computing Research Conference
Received: from Neon.Stanford.EDU by SAIL.Stanford.EDU with TCP; 9 Oct 89 18:52:49 PDT
Received: from Polya.Stanford.EDU by Neon.Stanford.EDU with SMTP (5.61/25-eef) id AA19868; Mon, 9 Oct 89 18:53:32 -0700
Received: from vm1.NoDak.edu by Polya.Stanford.EDU with SMTP (5.61/25-eef) id AA22732; Mon, 9 Oct 89 18:49:24 -0700
Message-Id: <8910100149.AA22732@Polya.Stanford.EDU>
Received: from NDSUVM1.BITNET by VM1.NoDak.EDU (IBM VM SMTP R1.2.1MX) with BSMTP id 1712; Mon, 09 Oct 89 20:46:45 CDT
Received: from NDSUVM1 by NDSUVM1.BITNET (Mailer R2.03B) with BSMTP id 1673;
Mon, 09 Oct 89 20:46:38 CDT
Date: Mon, 9 Oct 89 20:43:18 CDT
Reply-To: Theory-A - TheoryNet World-Wide Events <THEORY-A@VM1.NoDak.EDU>,
Joseph Traub <traub@cs.columbia.edu>
Sender: TheoryNet List <THEORYNT@VM1.NoDak.EDU>
Comments: Warning -- original Sender: tag was THEORYNT@YKTVMX
Comments: Warning -- original Sender: tag was THEORY-A@NDSUVM1
From: Joseph Traub <traub%CS.COLUMBIA.EDU@VM1.NoDak.EDU>
Subject: Strategic Directions in Computing Research Conference
To: Multiple recipients of list THEORYNT <THEORYNT@NDSUVM1>
The Conference on Strategic Directions in Computing Research,
sponsored by the Association for Computing Machinery (ACM) and the
Computing Research Board (which represents the Ph.D. granting
computing science departments), takes place in Arlington, Virginia on
October 11 - 13, 1989. The purpose of the conference is to review
current research issues in computing, to critique existing programs,
and to discuss directions and priorities. The conference is organized
into six topical panel sessions, each centered on an existing or soon
to be released report on the topic and on major government
initiatives. Each panel will be composed of the report author, a
Government program manager, 3-4 distinguished scientists, and a
session rapporteur. Following is a copy of the program and
registration information.
ACM/CRB Conference on Strategic Directions in Computing Research
Final Program
WEDNESDAY, October 11, 1989
7:00 p.m. - 8:30 p.m. Conference Context
This session will overview the objectives of the Conference. Particular
emphasis will be placed on overarching issues, including growth of
the discipline, Federal funding patterns, the taxonomy of the field,
priorities and balance in research activities, mechanisms for
educating and informing the public, the role and responsibilities
of professional societies and the need for a National forum on
computing research policy and issues.
Chairs/Organizers: David Gries, Paul Young
Speakers: David Gries-Cornell/CRB, Paul Young-Washington/CRB, Peter
Denning-RIACS/ACM, John White-Xerox/ACM
8:30 p.m. - 10:00 p.m. Reception with light food
THURSDAY, October 12, 1989
9:00 a.m. - 10:00 a.m. Welcome and Keynotes
David Wood, General Chair, Richards Adrion, Program Chair
Keynote Addresses: Congressman Bill Green
10:30 a.m. - 12:15 p.m. High Performance Architectures
High performance computing has been the subject of a series of
Federal Coordinating Committee on Science, Engineering and Technology
(FCCSET) reports of the Office of Science and Technology Policy (OSTP).
Major Federal programs of research, such as the DARPA Strategic
Computing Initiative and the NSF Computational Science Initiative,
and infrastructure support, such as the NSF Supercomputing Centers
Program, have been established in response to these reports. This
session will focus on many issues including: access to advanced
computing resources (centers versus high-performance workstations);
software and systems support; computing research and its role in
computational science; priorities and trends in architectures for
development of high-performance computing systems (distributed,
vector and parallel architecures, tightly- versus loosely-coupled,
special- versus general-purpose); and strategies for maintaining
and improving our competitive position in advanced computing.
Moderator/Chairman: J. C. Browne-Texas
Speakers: Steve Squires-DARPA, Gordon Bell-Ardent, Ken Kennedy-Rice
Larry Smarr-Illinois, Charles Seitz-CalTech
Rapporteur: J. C. Browne
1:30 p.m. - 3:15 p.m. Algorithms and Theory
In a field as new as computing research, theoretical studies, while
underway since the beginning of the field, still face major
challenges. As noted in the National Research Council report
on the National Challenge in Computer Science and Technology, research
on algorithms is not only of fundamental importance to the field, but
has resulted in a wealth of important, practical algorithms. This
session will focus on trends in theory and algorithms research, the
need for substantial and stable support of the subdiscipline, the
role of theory in computing research, and the unique character
of theory in a science based on evolving technology.
Moderator/Chairman: A. V. Aho-ATT/BL
Speakers: Richard DeMillo-NSF, Robert Tarjan-Princeton, Albert
Meyer-MIT, Richard Karp-UCB, Juris Hartmanis-Cornell, Joseph
Traub-Columbia
Rapporteur: Arnold Rosenberg-UMass
3:45 p.m. - 5:30 p.m. Artificial Intelligence
The subdiscipline of artificial intelligence has received considerable
attention over the last several years. AI technology is widely used
in commercial systems. AI has been the subject of numerous Government
reports and many significant initiatives, particularly in the Dept.
of Defense, have been established. This session will concentrate on
three areas in which major opportunities exist: the role of artificial
intelligence in scientific discovery; management of very large
knowledge bases; and intelligent learning environments. Recent advances
in these areas will be discussed and strategies for achieving rapid
advances in these areas will be proposed.
Moderator/Chairman: Raj Reddy-CMU
Speakers: Y. T. Chien-NSF, Douglas Lenat-MCC, Herb
Simon-CMU
Rapporteur: Raj Reddy-CMU
Reception/Banquet.
Speaker: William Poduska-Stellar Computer
FRIDAY, October 13, 1989
8:30 a.m. - 10:15 a.m. Systems and Software
In the evolving and complex systems upon which society depends
for financial management, social services, transportation, energy
delivery, production and defense, computers are beginning to
play a central and critical role. Advances in software and systems
have not kept pace with the rapid development of computer hardware.
Yet, it is the software and systems where the chief liklihood of error
exists and where the greatest development costs occur. This session
will analyze the challenges in software technology, evaluate existing
and proposed research programs, and sugest a strategy for the U.S.
in this most important area.
Moderator/Chairman: Vic Basili-Maryland
Speakers: William Scherlis-DARPA, Susan Gerhart-MCC, Lee Osterweil-UCI
Rapporteur: Mary Shaw-CMU
10:45 a.m. - 12:30 p.m. Neural Networks
Modern computers do not simulate biological systems in the way
they compute, analyze and store information. Recently, research
in neural networks has gained substantial National attention as
a significantly different approach to computation. The subject of
a major, forthcoming DARPA report and the focus of several Federal
programs, neural network research is the subject of this session.
Research in this area draws on psychology, linguisitics, neuroscience,
computation, cognitive science and mathematics. The panel will
discuss trends and advances in the field, the practical import
of neural networks, and strategies for the next decade.
Moderator/Chairman: Michael Arbib-USC
Speakers: Barbara Yoon-DARPA, David Rummelhart-Stanford,
Jacob Barhen-JPL, Dan Hammerstrom-Adaptive Systems
Rapporteur: Michael Arbib
1:45 p.m. - 3:30 p.m. Technology to Support Computer-Based Collaboration
Computer networks have become extremely important to science
and engineering research. However, they merely indicate the
possibilities for geographically distributed collaboration in
scientific research. This session will discuss a recent NSF
workshop on technologies and techniques to enable such collaboration
with an emphasis on access to information on the scale of National
online libraries, intelligent user support technologies, intelligent
instruments and the potential of telescience.
Moderator/Chairman: William Wulf-NSF
Speakers: Mark Stefik-Xerox,
Barry Leiner-RIACS, Robert Kahn-NRI, Lee Sproull-CMU
Rapporteur: Vinton Cerf-NRI
3:30 p.m. - 4:30 p.m. Wrap-Up
The conference will be summarized with emphasis on plans, directions,
and priorities.
Chairman: Richards Adrion-UMass, David Wood-MITRE
Speakers: Michael Arbib-USC, J. C. Browne-Texas, Vinton Cerf-NRI,
Raj Reddy-CMU, Arnold Rosenberg-UMass, Mary Shaw-CMU, Paul
Young-Washington
4:30 p.m. Adjourn
REGISTRATION INFORMATION
Fees:
ACM member or member of CRB organization $395
Non Member $495
Registration fee includes admission to technical sessions,
coffee breaks,luncheons,conference banquet and a post conference
report.
Make check payable to: ACM/CRB Strategic Directions
Send registration with payment to:
ACM/CRB Strategic Directions
ACM Conference Department
11 West 42nd Street
New York, NY 10036
(212) 869-7440
FAX: (212) 302-5826
e-mail: meetings @ acmvm.bitnet
******
FEARS ABOUT THE CONFERENCE
Preparing this conference was clearly a great deal of work.
To undertake their share of this work people had to be strongly
motivated to plan
``Strategic Directions in Computing Research'',
i.e. to plan other people's research. Those of us who
are too busy doing research or planning their own research
are going to find themselves with unpleasant choices.
1. Go to the conference and similar conferences and fight it out
with the people who have nothing better to do and get some
recognition of their own ideas. This won't be easy, because
victory in politics is like victory in any other field; it
goes to the people who are most professional at the field.
Here the field is politics.
2. Find themselves competing for a diminishing nominal dollar
pool devoted to peer reviewed research proposals in computer
science in general.
3. Carefully read the ``research initiatives'' resulting from
the round of conferences and either do what the planners
dictate or find some weasel words that will let them sneak
under the tent.
When large facilities are required for research, planning and
priorities are essential. For example, the particle physicists
had to reach an agreement on the Superconducting Super Collider,
and the astronomers have to reach agreement on new telescopes
and new spacecraft. The particle physicists and astronomers
have to accept the limitations on individual initiative that
such giant facilities enforce. They also have to accept the
limitation that good organizers will determine the direction of
research.
Such limitations are unnecessary and harmful in most of computer
science including AI. When I started in AI at Stanford in 1962,
a suitable computer for on-line AI research cost as much as 30
man years of researcher time. A large project with many people
was required to obtain the facility. Now most AI research can be
done with computers costing less than one man year of researcher
time. Only a small part of computer science requires planning
on the scale proposed here.
None of the work that I have done or plan to do in the near
future will easily fit under the three topics listed for
AI. They are all oriented towards applications and
pseudo-applications.
I see nothing in the conference that suggests any plan to
reverse the shrinkage in recent years of NSF support for
peer reviewed proposals unrestricted as to area of computer
science in general or area of AI in general.