ARMS-D-Request@XX.LCS.MIT.EDU (Moderator) (10/09/86)
Arms-Discussion Digest Thursday, October 9, 1986 9:15AM
Volume 7, Issue 28
Today's Topics:
Trump Cards vs. Broad Brushes
defenses against nuclear weapons
Viking Mars wonderfully successful, but slightly moot to SDI question
SDIO plans for cruises and bombers.
bias
Phil and SDI
advance warning of Libya attack
"We have only begun to research the ICBM defense problem"??
surrogate weapons
knowledge and being co-opted...
Autonomous Weapons
Strategic Deception
----------------------------------------------------------------------
Date: Mon, 6 Oct 86 07:36:00 PDT
From: Clifford Johnson <GA.CJJ@Forsythe.Stanford.Edu>
Subject: Trump Cards vs. Broad Brushes
> The insiders always have that trump card in any argument: "Classified
> information that you haven't seen and I can't tell you about proves
> you're wrong."
>
> I disagree that this is always a trump card. Details are classified,
> but broad brush strokes are not. The really fundamental arguments
> turn on broad brush strokes.
Then please tell me how I can find out whether the President has
contingency-predelegated the use of nuclear weapons, and, in
broad-brush terms, to whom and under what contingencies?
To: ARMS-D@XX.LCS.MIT.EDU
------------------------------
Date: Mon, 6 Oct 86 09:25:59 pdt
From: Gary Chapman <chapman@russell.stanford.edu>
Subject: defenses against nuclear weapons
Robert Maas says that there probably aren't any programs going on for
intercepting nuclear delivery methods besides ICBMs. In fact there
are lots of other R&D programs investigating the interdiction of
cruise missiles, depressed trajectory SLBMs, tactical missiles, etc.
The United States and several NATO countries are hard at work
developing a theater-based anti-tactical ballistic missile defense.
And according to an article published in the San Jose Mercury News,
President Reagan was actually going to announce a program to intercept
cruise missiles and SLBMs in the same speech in which he announced the
SDI, but he was talked out of it at the last minute by, of all people,
Richard Perle. (Conservatives have long maintained that the Soviets
are developing an anti- cruise missile and SLBM capability with the
upgrade of their already extensive air defense system, and our
conservatives have continuously complained about this country's lack
of an air defense system. Former Secretary of Defense James
Schlesinger testified before Congress that if the U.S. were to build
an air defense against cruise missiles it would cost at least $50
billion, and wo would cost $1 billion a year just to maintain.)
-- Gary Chapman
------------------------------
Date: 1986 October 05 13:47:47 PST (=GMT-8hr)
From: Robert Elton Maas <REM%IMSSS@SU-AI.ARPA>
Subject:Viking Mars wonderfully successful, but slightly moot to SDI question
DB> Date: Wed, 24 Sep 86 18:01:49 pdt
DB> From: Dave Benson <benson%wsu.csnet@CSNET-RELAY.ARPA>
DB> Subject: Viking Landers worked the first time and met the specs.
DB> Both Viking Landers worked in their first (and only) operation. The
DB> pre-operation testing simply ups one's confidence that the actual
DB> operation will be successful. Since the Viking Landers were the first
DB> man-made objects to land on Mars, Murphy's Law should suggest to any
DB> engineer that perhaps something might have been overlooked. In actual
DB> operation, nothing was.
Yup, only extreme care to look at possible bugs over and over again, and
a little luck, resulted in total success. It shows that in a benign
(non-hostile) and tame (non-random) environment it really is within the
state of the art (and was in 1974) to defeat Murphy once. But try the
same trick on Earth where a hailstorm or small animal can accidently
knock over the equipment and spoil the whole mission, and the chance of a
totally successful mission becomes more remote. Or try an actively
hostile environment such as the streets of Harlem or East Palo Alto where
some "turkey" is sure to vandalize the equipment unless an armed guard is
posted to protect it (which of course defeats the purpose of the robot
probe to alien territory where no human can yet go). Even Venus, which is
moderately unpleasant/random, has presented some problems to the USSR
craft which have landed there. Still, an active enemy trying to break
your program is far harder to handle.
Note also the Viking program had no judgement or other A.I., just
interpreting commands from Earth and dutifully executing them in a
mechanical way. A.I. programs are much less likely to work perfectly for
years the first time they are really tried out of the test environment
(my opinion).
I would say this example shows a lower bound on what is possible, but
it's far from what is needed for SDI. We need to have a successful
mission under actually hostile circumstances before we can really apply
the example directly to say SDI might be possible.
DB> Both Viking Mars shots had specifications for the length of time they
DB> were to remain in operation. While I do not recall the time span,
DB> both exceeded the specification by years.
Once you get a program working, it generally works just about forever
until you try to use it for something new (antenna not aimed at Earth,
OOPS!!) or something physical breaks down (cosmic ray etc.) or some
hostile agent figures out how to defeat it (in hostile situation).
Generally our spacecraft have exceeded their planned lifetimes if they
worked at all, thus Viking doesn't surprise me given that it worked at all.
DB> Surely any engineered artifact which lasts for longer than its
DB> design specification must be considered a success. Nothing
DB> lasts forever, especially that most fragile of all artifacts, software.
DB> Thus the fact that the Viking 1 Lander software was scrambled beyond
DB> recovery some 8 years after the Mars landing only reminds one that
DB> the software is one of the components of an artifact likely to fail.
DB> So I see nothing remarkable about this event, nor does it in any way
DB> detract from judging both Viking Mars missions as unqualified engineering
DB> successes.
I agree completely. Viking was a glorious success within its task design.
Too bad the rover never came as a follow-up, leaving Viking as a dead end
for over ten years. (P.s. those pictures of the Dunes of Mars were really
pretty, like some quiet Earth desert near a river where I might play as a
child. I sort of wish some place like that existed around here for me to
relax in every so often. Anyway, the people who predicted Mars would have
dunes were right.)
------------------------------
Date: Mon 6 Oct 86 16:11:08-EDT
From: Marc Vilain <MVILAIN@G.BBN.COM>
Subject: SDIO plans for cruises and bombers.
When General Abrahamson spoke in Boston last October (at the Ford
Hall Forum), he was asked what the SDIO planned to do about cruise
missiles, bombers, and "suitcase bombs". His response, predictably, was
that the SDIO wasn't ignoring the problem. But, he said, SDIO believed
that the problem of intercepting ICBMs was much more difficult than that
of intercepting cruise missiles and bombers, and hence should be
addressed first. The other problems would naturally be conquered in
turn.
With respect to specifics, Abrahamson pointed to current R & D
programs (outside of SDIO) as providing steps towards solutions of the
cruise/bomber problem. On the sensor side, he mentioned work towards
detecting turbofan engines (even small ones) from space -- I assume he
meant TEAL RUBY here. On the interceptor side, he mentioned nothing, as
I recall.
Once again, this was last October, and the SDIO may have a different
position today.
marc vilain (mvilain@g.bbn.com)
------------------------------
From: AI.DUFFY@R20.UTEXAS.EDU
To: Richard A. Cowan <COWAN@XX.LCS.MIT.EDU>
Cc: arms-d, lkk@XX.LCS.MIT.EDU, prog-d%OZ.AI.MIT.EDU@Reagan
Subject: bias
[Forwarded by Cowan]
Date: Thursday, 2 October 1986 00:15-CDT
From: Richard A. Cowan <COWAN at XX.LCS.MIT.EDU>
While on the subject of bias in media, an exchange I had with
Herb Lin recently may be relevant:
[cowan@xx]
US system may be more democratic, but the US population is
also controlled, when you consider how the mass media
(especially television!) narrows the debate and ratifies the
existing distribution of power by trying hard to avoid giving
credence to "controversial" positions.
[lin@xx]
My limited experience with electronic mass media (having been
interviewed twice for broadcast) is that they go out of the way to
accomodate controversial positions. Indeed, my criticism is that they
have tried to polarize the debate even MORE than is justified. They
were reluctant (but ultimately willing) to accept all of my
qualifiers, looking instead for journalistic "punch".
Comments?
Speaking from a far less limited experience with the mass media, you
are both correct.
The electronic mass media does indeed encourage controversy, but only
particular kinds of controversy. Television is particularly fond of
promoting controversy that includes action photography and what Lin
calls "journalistic punch". They are less fond of conceptual
controversy -- partly because this doesn't make for good pictures and
partly because they're trying to reach the lowest common denominator
in the potential audience (I think "journalistic punch" means content
that the lowest common denominator can comprehend).
For a discussion of some of the relevant issues, I humbly suggest my
essay, "The Normative Ground of Spectrum Policy Debates", in Brenda
Dervin and Mel Voigt, eds., _Progress_in_Communication_Sciences_, v.
7 (Ablex, 1985 or 1986).
------------------------------
Date: Mon 6 Oct 86 20:25:15-EDT
From: Richard A. Cowan <COWAN@XX.LCS.MIT.EDU>
Subject: Re: Re: Phil and SDI
; From: Richard A. Cowan <COWAN>
;
; ; There are also quite a large number of people who think [SDI] is the
; ; only morally and ethically concievable way of defending ourselves.
;
; There is quite a large number of people who will make a lot of money if
; SDI funding continues. It is only human nature for these people to
; justify what they do by claiming it will defend "us."
;
; From: Herb Lin <LIN>
; These two aren't inconsistent. It's a red herring to claim that the
; only ones who think the U.S. should buy weapons are those who will
; profit from it.
You are wildly exaggerating my argument. I do not claim that the ONLY
people who think the US should buy weapons are the ones that profit
from it. Plus, I was specifically referring to SDI, which has
relatively little support (compared to other weapons systems) in the
academic community except for those who have been "bought by the
bucks."
; ; There has got to be a better way to protect our right to be left
; ; alone, and it is worth trying to make it real.
;
; Finally, the idea that the US is merely trying to be "left alone" and is
; leaving the affairs of other countries alone is also absurd.
; ... There is a better way to
; protect our right to be left alone, and that is to leave others alone!
;
; These two statements are not inconsistent. We do need military force
; to protect our right to be left alone, and we also should not use
; military force to the extent that we do to bother others. Force
; should be the option of last resort, not the option of first resort
; and not an unacceptable option.
Again, you are characterizing me as absolutely opposed to any military
force. That is not so; I agree that there are situations for which
the US military serves a legitimate defensive purpose. I should
mention that it is not just the USE of military force that I was
talking about, but also the THREAT of its use, which has bothered
others (I am not talking about the USSR) by enabling the US to enter
into agreements with those countries that are more favorable to us
than would be the case without our military edge.
-rich
------------------------------
From: decvax!utzoo!henry@ucbvax.Berkeley.EDU
Date: Mon, 6 Oct 86 23:42:50 edt
Subject: advance warning of Libya attack
I happened to be catching up on old issues of Flight International, and
was struck by the following report, published in the "World News" section
just *before* the attack on Libya:
"The United States Air Force (Europe) was alerted on April 11, it is
believed, in readiness for a combined strike against Libya together with
the US Navy.
"The alert was signalled by a massive influx of tanker aircraft into the
UK, with 16 McDonnell Douglas KC-10 Extenders arriving at RAF Mildenhall,
and another eight KC-10s going to RAF Fairford. Mildenhall's usual
complement of Boeing KC-135 Stratotankers was up from 14 aircraft to 20,
including an extra KC-135Q, used exclusively to refuel Lockheed SR-71s
which are also based at Mildenhall.
"USAF(E) says that the unusual activity is in preparation for a Nato
exercise, but *Flight* understands that Mildenhall's Sunday morning curfew
on flying was broken for the first time, with about eight KC-10s arriving
between 0200 and 1000 on Sunday April 13. The F-111s of the 48th Tactical
Fighter Wing were being ground-run all Saturday evening.
"The use of KC-10s would give Pentagon planners greater flexibility as
they are capable of refuelling both US Navy probe-equipped aircraft and
of using the boom-to-refuel receptacle fitted USAF aircraft.
"* As we go to press, an airline pilot reports seeing two aircraft carriers
and 'lots of smaller ships' heading 150-160 deg. from Sicily towards the
Gulf of Sirte on the Libyan coast."
(Flight International, issue dated 19 April 1986, page 2.)
Henry Spencer @ U of Toronto Zoology
{allegra,ihnp4,decvax,pyramid}!utzoo!henry
------------------------------
Date: Tue, 7 Oct 86 11:50:50 PDT
From: jon@june.cs.washington.edu (Jon Jacky)
Subject: "We have only begun to research the ICBM defense problem"??!!
> (Robert Elton Maas writes) It may be too early for anyone to make plans or
> solid cost estimates, considering we have only begun to research the ICBM
> defense problem and haven't started any research at all on other defense...
We have NOT "only begun" to research ICBM and other strategic defense. The US
has been doing ICBM defense research since the mid-1950's. From time to time
it erupts into a big deal, thanks to periodic bursts of presidential attention
and public interest. The current Star Wars/SDI hoopla is only the most recent
eruption; the Safeguard/ABM business in the 1960's was another biggie. Note that
US spending on missile defense was around $1 billion / year before SDI (its
about $3B/yr now). Note also that the annual budget for the entire NSF is
also about $1B/year, and the annual budget for the National Cancer Institute
is about $1B/year as well. It is seldom argued that the US is making an
insignificant research effort in these realms.
As for other defense, presumably meaning defense against bombers, cruise missiles,
and other airbreathing threats: conventional air defense has probably
consumed a considerable fraction of the world's technical efforts since 1914.
-Jonathan Jacky
University of Washington
------------------------------
Date: Tue, 7 Oct 86 20:34:44 PDT
From: lai@ucbarpa.Berkeley.EDU (Nick Lai)
Subject: surrogate weapons
The goal of the recent discussion of autonomous weapons appears to
be to decry their existence, and to call upon all decent people to
work towards banning them.
While the thought of autonomous weapons, say a robot tank, roving
the countryside, killing civilians, or human military for that matter,
is unsettling, it also seems unrealistic. In order to build such
weapons, one would need a very sophisticated real-time pattern analysis
and recognition mechanism, as well as a decision system which could
make "intelligent" decisions from the information coming from these
devices.
From research accounts I've read in the pattern analysis field,
digesting "real-life" pictures into intelligible data is a
very difficult problem. For an example, see [Burns, 1986],
which describes an algorithm for extracting straight lines
from a digitized "real-life" image (attempting recognition on
a raw data frame of 1000 x 1000 = 1000000 pixels is unadvisable).
The operation is computationally very expensive, and does not seem
feasible on a real-time scale. Further, as one makes the algorithm
more discriminatory to get a smaller number of better lines,
the results are harder to recognize as the original object to
a human, let alone a computer.
Once one had the extracted lines, one could develop 2D to 3D geometric
matching algorithms to classify things as "house", "tank", "bridge",
etc., in simple cases. But "real-life" is complex, with obscuring objects,
observation angles, camouflage, depth, and a great number of variations on
simple structural themes.
While I think that truly autonomous weapons are unrealistic,
I believe that we are going to see more and more of what I
will call "surrogate weapons", which I define as a self-propelled
weapon which is guided remotely by a human. A wire-guided
torpedo is a trivial example of this. However, I am thinking more
along the lines of a tank, aircraft, or submarine which
is controlled by a human from a greatly removed position. The
communication between human and weapon would be full-duplex,
with audio/video/sensor information being transmitted to
the human master, who digests the data and sends back
commands to the remote weapon.
The reason I call it a "surrogate" instead of "remote" weapon
is that I consider the weapon to be "standing in" for the
human who can direct destruction from a safe, comfortable
distance.
These weapons will be considerably easier to build than
autonomous weapons, since they require no artificial vision
or intelligence, and are likely to be far more reliable.
In addition, they share the prime "benefit" of autonomous weapons:
no white-hat combatives get killed.
Are these surrogate weapons "immoral" or otherwise undesirable?
It is open to debate.
As was brought up in the autonomous weapons discussion,
parties in possession of these weapons are more likely
to use them. Americans are very hesitant about committing
US troops to various parts of the world, after such traumatic
experiences as Vietnam and the loss of the Marines in the
Lebanon bomb attack. I suspect, however, that the general
population could not care less if we stationed surrogate
battle-tanks in El Salvador which were controlled by marines and
video-game junkies stationed in Iowa.
Further, I believe that people who are engaged in a battle in which
their lives are at stake develop a certain amount of compassion
for their opponents. From a distance of ten miles or a thousand
miles, humans may be less inclined to feel compassionate.
In any event, I am curious if anyone on this list knows of any
project to develop the kind of weapons described above, and
invite discussion on their feasibility / acceptability.
Nick
References:
[Burns 86] J. B. Burns, A. R. Hanson, and E. M. Riseman, "Extracting
Straight Lines," IEEE Transactions on Pattern Analysis and
Machine Intelligence, Vol. 8, No. 4, pp. 425-455, 1986.
------------------------------
Date: Wed, 8 Oct 1986 01:54 EDT
From: Rob Austein <SRA@XX.LCS.MIT.EDU>
Subject: knowledge and being co-opted...
Date: Wednesday, 1 October 1986 17:46-EDT
From: Hoffman.es@Xerox.COM
No, but I think it's IMPOSSIBLE for an outsider to know ENOUGH about
matters related to defense to be an effective part of the solution.
The insiders always have that trump card in any argument: "Classified
information that you haven't seen and I can't tell you about proves
you're wrong."
Maybe that's a winning argument in your book. It sure ain't one in
mine. Given the kind of idiotic uses that classification has been put
to in this country (and all others, but this one is supposedly a
democracy), I am completely unable to tell whether you are telling the
truth or are making this up (to win a multi-million dollar contract or
to save face or to save your boss's face or ...) [*].
So, if I am trying to make an informed decision (as a responsible
citizen in a democratic country) your alledged inside information is
totally useless to me. I have to make my decision without the benefit
of your information. To be sure, I may decide wrong, but it's the
best I can do. If you want me to decide correctly, then show me your
inside information. But don't expect me to stop voting or stop
writing my congress-critter or stop rabble-rousing or support
technical decisions which I don't think are correct on account of
information you refuse to show me.
("You" in the above does not refer to Mr. Hoffman except in the first
sentence, of course.)
--Rob Austein <sra@xx.lcs.mit.edu>
[*] For a case study of idiotic uses of classification by the US
government, see "The CIA and the Cult of Intelligence" by
Victor Marchetti and John D. Marks (Dell, New York, 1974 &
1980). Some of the original information (regarding the CIA)
is of less interest than when the book was first published,
but the history of the censorship of that book is absolutely
facinating. The copy I have indicates where (and how much)
material was removed by the censors, as well as indicating
portions which the CIA tried to have censored but which were
allowed back in after court battles. If you think there is no
such thing as goverment censorship in this country you had
better read this book. --sra
------------------------------
Date: Monday, 6 October 1986 16:07-EDT
From: convex!paulk at a.cs.uiuc.edu (Paul Kalapathy)
To: ARMS-D
Re: Autonomous Weapons
It seems to me that for a more useful definition of autonomous
weapons, that one should either consider ACTIVE autonomous weapons,
or that non-active autonomous weapons should be disregarded as a
degenerate case.
As a strawman definition, I would consider an active autonomous
weapon to be a weapon which is:
1) capable of seeking and selecting (without the immediate
intervention of some person) a target from a field
of inputs which could contain targets, non-targets,
neither or both.
2) capable of altering behavior (beyond the act of 'going off')
in order to destroy the target which it has selected without
the immediate intervention of some person.
I think the discussion of non-active autonomous weapons has
bogged down the discussion. Consideration of land mines as autonomous
weapons is relevent in only the most obtuse way. If a land mine is
considered an autonomous weapon, then surely the tiger trap (a hole
in the ground with some big leaves over it) is also an autonomous
weapon. The land mine explodes when some pressure is put on it, and
the leaves on a tiger trap collapse when some pressure is put on them
causing the poor beast or person to fall in (onto pointy stakes if
you prefer).
The consideration of neutron bombs as autonomous weapons is also
useless except as a trivial case. If neutron bombs are AW, then so
are tear gas, chemical and biological agents, and the destruction of
crops which starves only the living.
The above definition would exclude mines and tiger traps since
they do not discriminate between targets and non-targets, and do not
alter their behavior based on inputs (other than to 'go off').
Neutron bombs and their ilk are excluded since they are
non-discriminatory; neutron bombs irradiate everything, its just that
people don't take as kindly to that as do bricks. Also excluded are
weapons which are pointed at some target and then use some guidance
system to home in, since they require the immediate intervention of
some person in order to select the target (e.g., anti-aircraft missiles,
anti-ship missiles, etc.)
The above definition would include such AWs as fire-and-forget
missles which are supposed to select tanks or whatever from a
battlefield that they are tossed at, autonomous land vehicles that are
supposed to drive around and blow up tanks, and other weapons which
seek a target without the immediate assistance of some person.
Paul Kalapathy
CONVEX Computer Corp.
Richardson, TX
------------------------------
Date: 9 Oct 1986 08:21 EDT (Thu)
From: Wayne McGuire <Wayne%OZ.AI.MIT.EDU@XX.LCS.MIT.EDU>
Subject: Strategic Deception
Four quick observations and two questions:
(1) We have just learned that the Reagan Administration has engaged in
a policy of so-called disinformation or "strategic deception" regarding
Libya and calculated to rattle Khadafy. While American intelligence data
revealed that Khadafy's terrorist planning activities were in a "quiescent"
phase, the Administration falsely claimed precisely the opposite, that
Libya was hatching new terrorist schemes.
(2) Seymour Hersh in _The Target Is Destroyed_ has demonstrated quite
convincingly, with the cooperation of disgruntled members of the American
intelligence community who were appalled by the abuse of intelligence data,
that the Administration knew clearly that the Soviet Union mistakenly
thought that KAL 007 was a spy plane. Again, the Administration in the
interests of pursuing an ideological offensive, turned the truth upside
down to score a few propaganda points in charging that the Soviets
deliberately and knowingly attacked a civilian airliner.
(3) A few years ago hysterical stories, supposedly based on classified,
inside information, appeared in the American media about the sinister
presence in the U.S. of a Libyan "hit team." Later more level-headed
information indicated that the story was a fantasy and probably cooked up
by Israeli intelligence as a means to stir up fear and hatred of Khadafy,
and to aggravate tensions between the U.S. and the Arab world.
(4) James Bamford, author of _The Puzzle Palace_, a popular study of
the National Security Agency, recently commented in _The Boston Globe_ that
the Administration seriously compromised intelligence methods by providing
details about how communications were intercepted pertaining to the
terrorist bombing of a discotheque in Germany, the proximate cause of our
bombing of Tripoli, but failed to release the content of those
communications so that objective analysts could determine whether they did
indeed implicate without a doubt the Libyan government in this terrorist
incident.
Question 1: doesn't one begin to see a fairly consistent pattern of
deception here, and doesn't it raise some serious questions about what is
its purpose, who benefits, and what is a judicious use of intelligence
information in policy-making?
Question 2: I am quite willing to believe that Khadafy is the arch-
terrorist fiend and monster that the media have painted, but whenever I
have asked some of the people who seem most upset by this problem to
produce hard evidence that the Libyan government has engaged in terrorism
against American citizens, at a level that would justify the bombing of
Tripoli, I have encountered a good deal of emotional language but no clear
facts. I am eager to be enlightened by anyone on the list who does possess
any facts: specifically, what American citizens during the last decade have
been the targets of terrorist attacks by the Libyan government or its
surrogates? Names and particulars, please.
------------------------------
End of Arms-Discussion Digest
*****************************