NetNews Usenet Archive 1992 #20/NN_1992_20.iso/spool/comp/theory/1877

Text File | 1992-09-08 | 17.2 KB | 401 lines

Newsgroups: comp.theory
Path: sparky!uunet!stanford.edu!CSD-NewsHost.Stanford.EDU!CSD-NewsHost!jmc
From: [e-mail] (John McCarthy)
Subject: Petition for withdrawal of NRC report Computing the Future
Message-ID: <[e-mail]>
Sender: [e-mail]
Reply-To: [e-mail]
Organization: Computer Science Department, Stanford University
Date: 5 Sep 92 22:01:46
Lines: 389

  A petition sponsored by John McCarthy, Bob Boyer, Jack Minker,
  John Mitchell, and Nils Nilsson.

Dear Colleagues in Computer Science and Engineering:

We are asking you to join us in asking the Computer Science and
Telecommunications Board of the National Research Council to withdraw
for revision its report entitled ``Computing the Future: A Broader
Agenda for Computer Science and Engineering'', because we consider it
misleading and even harmful as an agenda for future research.  Our
objections include its defining computer science in terms of a narrow
set of applied objectives, and its implication that the tone of
computer science is to be set by government agencies, university
administrators and industrialists and that computer scientists are
just the ``soldiers on the ground''.

There is much useful information in the report, but the Preface and
the Executive Summary characterize computer science in a way that no
other science would accept.  Chapter 2, ``Looking to the Future of
CS&E'', and Chapter 3, ``A Core CS&E Research Agenda for the Future''
should also not be accepted by computer scientists.  The Report merges
computer science and computer engineering at the cost of abolishing
computer science and seriously narrowing computer engineering.

Besides individual scientists, we hope that some computer science
departments will collectively join in requesting the report's
withdrawal.

Our campaign for the report's withdrawal is being conducted entirely
by electronic mail, and we will be grateful to anyone who forwards
this message to others who might be concerned.  Email to
[e-mail] will be counted as signing the following
petition, not as necessarily agreeing to everything in this message.
In fact, the sponsors of this message are committed to the petition
and not necessarily to every detail of the message.  We haven't taken
the time to hash out every detail.  So ``sign'' if you endorse the
petition.

To read more before signing, see the end of this message.

This message contains the following parts:

 The introduction preceding this table of contents.
 The petition we are asking you to sign.
 The reasons why the report should be withdrawn.
 Some administrative details: email addresses, how to get the full report
    by regular mail, and how to get the preface and executive summary of
    the report by ftp.



                 THE PETITION

We, the undersigned computer scientists and engineers, find the report
``Computing the Future: A Broader Agenda for Computer Science and
Engineering'' misleading and even harmful as an agenda for future
research.  We have four objections.  First, it defines computer
science in terms of a narrow set of applied objectives.  Second, it
suggests that the tone of computer science is to be set by government
agencies, university administrators and industrialists and that
computer scientists are just the ``soldiers on the ground''.  Third,
it pre-emptively surrenders to the bad idea that basic research should
not be supported and that all support for science should be based on
``concrete demonstration of benefit to the nation.''  Fourth, it lacks
an adequate summary of scientific research goals and the resources
needed to achieve them.

We request that the Computer Science and Telecommunications Board
withdraw the report for rewriting, especially the Preface and
Executive Summary.  Chapters 2 and 3 also need to be replaced.  The
revision should discuss computer science in terms of the questions
computer scientists are curious about, and should emphasize that the
research goals, as in all sciences, are determined primarily by the
researchers.

End of Petition.  Email to [e-mail] will be
automatically counted as endorsing this petition.  Comments should go
to [e-mail].


        THE REASONS WHY THE REPORT SHOULD BE WITHDRAWN

1. What we most object to is exemplified by table ES.1 of the executive
summary and the note following it.

TABLE ES.1 Importance of Core Subfields of CS&E to Selected
                       Applications


                      Global Change Computational Commercial Electronic
Core Subfield         Research      Biology       Computing  Library

Multiple processors   Very         Central      Important  Very
                      important                              important

Data communications   Central       Important     Central    Central
 and networking

Software engineering  Important     Very          Central    Important
                                    important

Information storage   Central       Very          Very       Central
and management                      important     important

Reliability           Very          Important     Very       Important
                      important                   important

User interfaces       Very          Very          Central    Central
                      important     important


NOTE: The core subfields listed above constitute a FUTURE RESEARCH
AGENDA for CS&E. As significantly, they are important to, and can
derive inspiration and challenging problems from, these selected
application domains. The core subfields correspond to areas in which
major qualitative and quantitative changes of scale are expected.
These areas are processor capabilities and multiple-processor systems,
available bandwidth and connectivity for data communications and
networking, program size and complexity, management of large volumes
of data of diverse types and from diverse sources, and the number of
people using computers and networks. Understanding and managing
these changes of scale will pose many fundamental problems in CS&E,
and using these changes of scale properly will result in more powerful
computer systems that will have profound effects on all areas of human
endeavor.

[End of excerpt from Executive Summary.  The capitalization of FUTURE
RESEARCH AGENDA is ours.]

This simply doesn't do justice to computer science as a branch of
science.  Many, many areas of basic research are omitted, e.g.,
artificial intelligence and theoretical computer science.  On page 22
of the body of the report is a quite conventional taxonomy of computer
science taken from a paper by Peter Denning.  There the subfields are
listed as

 Algorithms and Data structures
 Programming languages
 Computer architecture
 Numeric and symbolic computation
 Operating systems
 Software engineering
 Databases and information retrieval
 Human-computer interaction

This is not great but is recognizable.  However, the Executive Summary
does not refer to this classification.

2. Our second objection is exemplified by the list on page viii of
addressees of the report.

     Given the increasing pervasiveness of computer-related
     technologies in all aspects of society, the committee
     believes that several key groups will benefit from an
     assessment of the state of academic CS&E:

      Federal policy makers, who have considerable influence in
     determining intellectual directions of the field through
     their control of research budgets and funding levels;

      Academic computer scientists and engineers, who are the
     ``troops on the ground'' that do research and teach
     students;

      University administrators, who play key roles in setting
     the intellectual tone of the academic environment; and

      Industry, which is by far the major employer of CS&E
     baccalaureate holders, one of the major employers of CS&E
     Ph.D.  recipients, and (in the computer industry) a key
     player in CS&E research.


     Each of these groups has a different perspective on the
     intellectual, fiscal, institutional, and cultural influences
     on the field, and the committee devoted considerable effort
     to forging a consensus on what should be done in the face of
     the different intellectual traditions that characterize
     various subfields of CS&E and of different views on the
     nature of the problems that the field faces.

The metaphor ``troops on the ground'' expresses an attitude toward
computer scientists.  The Report sees the content of computer science
as mainly determined by Federal policy makers, university
administrators and industry.  We ``troops on the ground'' are perhaps
to be asked for suggestions from time to time.  No report on the
research agenda of mathematics or physics would tolerate university
administrators ``setting the intellectual tone'' of the field.  It
wouldn't be imagined that university administrators had anything to do
with it.  The report is likely to encourage a regrettable bossiness in
otherwise perfectly reasonable deans.

One could argue that this is just an unfortunate wording that could
readily be fixed, howvever it fits well with the idea that computer
science has no independent existence but is determined by the problems
of the owners of computers.

3. Another dominant attitude of the report is expressed on page 3 by

     Assumptions of the 1940s and 1950s regarding the positive
     social utility of basic research (i.e., research without
     foreseeable application) are being questioned increasingly
     by the federal government, and justifications for research
     may well in the future require concrete demonstrations of
     positive benefit to the nation.

Indeed such tendencies exist, but in the week the report came out, the
President of the United States was in Texas drumming up support for
the Superconducting Super-collider, the most expensive device for pure
research ever to be built.  The report is based on a pre-emptive
surrender to anti-science tendencies that are only slightly worse now
than they have ever been.


                 DISCUSSION

These observations aren't arguments as to why the report should be
withdrawn but rather thoughts on how we got into this mess and what is
needed to develop ideas about what computer science is and how it can
be advanced.

Computer scientists have always had practical arguments as to
why our research should be supported.  This is unlike physics, where
for several centuries, the main justification had to be that it
improved mankind's understanding of the world.  Thus none of the
arguments for or against Galileo were concerned with whether his ideas
would help the renaissance papacy in its wars.  It has always been
easy to couch proposals for support of computer science in practical
terms.  Nevertheless, computer science does have its fundamental
problems and subfields.  It would have been well had the committee
tried to identify them; certainly many of its members have the
necessary qualifications.

Here is a try at identifying some of the problems that give computer
science its structure.  We can consider both history and current
problems.  It is just a sample, and we think the existing committee
could do a better job than this of identifying research goals if they
weren't diverted from thinking about science.  Anyway such a broad
committee is needed if the facilities and people requirements for
achieving these goals are to be comprehensively treated.

1. Consider the history of regular expressions.  They were
invented by the mathematical logician Stephen Kleene, and their first
appearance was in a RAND Corporation report about 1950.  In those days
RAND supported much basic research.  The motivation was to determine
what languages could be recognized by finite automata.  Kleene
presumably already knew that languages requiring that parentheses
match could not be recognized by finite automata, so natural language
recognition could not have been his objective.  The main application
of  regular expressions today is in string searching, completely
unanticipated by Kleene.  They were certainly not invented as part of
one of the areas in the Report's ``Core Research Agenda.''  If the
Report's recommendations were in force today, a proposal to study what
languages were recognizable by finite automata would lose out in
competition with proposals that more clearly provided ``concrete
demonstrations of positive benefit to the nation.''

2. McCarthy's work on Lisp and his work on time-sharing were both
motivated as means of carrying out his research on artificial
intelligence.

3. The relation between the facts that determine what an object is and
algorithms that compute it has been studied in mathematical logic and
in computer science (both in AI and in mathematical theory of
computation).  It is a permanent core research area of computer
science.  It can probably be related to all the core research areas
listed in the Report, but it will lose out in competition to much
narrower topics more immediately related to any of them.

4. What can be known about the world by a system with given
opportunities to observe and interact.  This is a key area
of artificial intelligence---the formalization of common sense.

5. How to express knowledge of the world in databases.

6. What are lower bounds on the number of operations needed
to perform various numerical and symbolic computations, e.g.
invert a matrix or store and retrieve information in
data structures?

7. What are the important data structures and the operations
on them.

8. What are the routine parts of a proof of a mathematical
fact, e.g. the correspondence of a computer program to its
specifications, that can be done routinely by computer, and
what are the creative parts that require either human attention
or substantial AI.

9. What are the properties of various search strategies.

10. What aspects of algorithms are parallel and what are essentially
serial?  This has a certain relation to the Report's multi-processors,
and undoubtedly someone could make a case for studying parallelism
in general on the grounds that it would help design or use
multiprocessors.  One could also make a case for studying
thermodynamics on the grounds that it is relevant to automobile
engines.  However, the committees that evaluate a physics proposal
will consider its importance for thermodynamics and not directly
for automobile engines.

11. Modularity.  What are the opportunities for modularity in
descriptions of the behavior of very complex systems?  This affects
both the analysis of existing systems and the synthesis of new
ones.

            EXPERIMENTAL COMPUTER SCIENCE

 Experimental computer science needs to be distinguished from
applied computer science.

 Some of it involves experimentally learning what algorithms embodying
certain data structures or certain heuristics or certain other ideas.

 Some of it involves programs that learn.

 Note that high quality displays arose in experimental computer
science long before workstations were proposed for engineering.

 The chess machines have taught us that certain heuristics are
necessary to reach human level performance no matter how much
computer power there is.  They have helped identify intellectual
mechanisms, and their present lavish use of computation tells
us that there is a lot still unknown about intellectual mechanisms.

 The experimental work on automatic and interactive theorem proving tells
us still more about what intellectual mechanisms we do and don't
yet understand.

 Experimental computer science may benefit from special equipment,
either special computers or auxiliary equipment.  There needs to
be a study of what facilities are required.  Fortunately, we are
unlikely to require equipment as costly as physics does.


               ADMINISTRATIVE MATTERS


To add your name to the signatories of the petition, send
email to [e-mail].  All mail to that address
will be counted automatically as supporting the petition.

Comments to the sponsors of the petition should be sent
to [e-mail] and will be distributed.
Comments directly to the Computer Science and Technology Board should
go to its Chairman William Wulf, [e-mail].

Copies of the full report can be obtained from

  National Academies Press
  2101 Constitution Ave.
  Washington, D.C. 20148

  1-800-624-6242

The price is $24.95 per copy + $3.00 shipping per order, and people
from California and Maryland have to pay sales tax.  Quantity
discounts exist.  Call them about it if you care.  Maybe you won't
want it for your permanent library, but if we are successful in
getting it withdrawn, maybe it will become a rare book.

[The Latex source of the foregoing arguments against the report
document is available by anonymous ftp from sail.stanford.edu under
the name /pub/jmc/whysign.tex.  The other two documents are
/pub/jmc/petition.tex for the petition itself and /pub/jmc/preface.tex
for the preface and executive summary of the NRC report.  A
non-Latex email message containing both the petition and arguments for it
may be found under the name /pub/jmc/petition-why.]

As of 1992 September 4, the sponsors of this request are
Bob Boyer, [e-mail],
John McCarthy, [e-mail],
Jack Minker, [e-mail],
John Mitchell, [e-mail]
and
Nils Nilsson, [e-mail].  John McCarthy may be telephoned
at 415 723-4430.







--
John McCarthy, Computer Science Department, Stanford, CA 94305
*
He who refuses to do arithmetic is doomed to talk nonsense.