ARMS-D-Request@XX.LCS.MIT.EDU (Moderator) (11/11/86)
Arms-Discussion Digest Monday, November 10, 1986 4:30PM
Volume 7, Issue 55
Today's Topics:
Administrivia: fee for service postings (PLEASE READ)
Re: killing someone if we had to look them in the eye
24 hour delay
24 hour delay
Nuke equivalnce/long endurance aircraft
Big R&D programs
Re: Military Institutions and AI
Re: SDI is impossible
portable anti-tank weapons
Yet more on SDI (Star Wars flawed #6-of-10)
----------------------------------------------------------------------
Date: Mon, 10 Nov 1986 16:18 EST
From: LIN@XX.LCS.MIT.EDU
Subject: Administrivia: fee for service postings...
I recently received for posting a note that involves a fee for service
arrangement. The service to be performed would be of interest to some
segments of the ARMS-D community: it described itself as a new
research and documentation service designed to provide up-to-date
detailed information about military contracting and companies involved
in military work. The organization providing the service is
non-profit and tax exempt, but it is charging money for the service.
Here is the dilemma. The June 1984 issue of the "DDN Directory" gives
on page 12 the following statement concerning use of the network:
"The DDN is an operational DoD network and is not intended
to compete with comparable commercial service. It is intended
to be used solely for the conduct of or in support of official
U.S. Government business."
It is the vagueness of the term "official U.S. Government business"
that forces me to make a judgment call on what activities constitute
"support of U.S. Government business." ARMS-D exists to discuss
issues related to war and peace, national security and the like, and
in doing so provides support for many individuals doing work for the
U.S. government. But what activities should be excluded?
It is clear that advertising for-profit activities or activities for
personal gain would be against this policy. But the non-profit nature
of the posting organization places the submitted note into a sort-of
limbo.
After pondering the question, I decided to ask the original poster to
resubmit an abbreviated note that leaves out explicit mention of any
cost involved, which I promised I would publish. In the future, I
will decline any submissions that involve fee for service
transactions.
As always, I welcome comments on whether or not I did the right thing.
Herb Lin
Moderator ARMS-D
------------------------------
Date: 10 Nov 86 10:03:19 EST (Monday) From: MJackson.Wbst@Xerox.COM
Subject: Re: killing someone if we had to look them in the eye
From: ihnp4!utzoo!henry@ucbvax.Berkeley.EDU (V7 #51):
"Soldiers do get desensitized to violence, but this doesn't turn them
into rabid killing machines, the apparent views of some liberals
notwithstanding. Combat-experienced infantrymen, at least from
Western cultures, are generally pacifists; patriotism and the glory of
war cut little ice with them. They fight because it's suicidal not to
in such an environment, and they stay in that environment because they
don't want to leave their buddies short-handed."
Basically correct, and not unknown to this liberal, at least. See
/The American Soldier: Combat and Its Aftermath/ (Stouffer, et al.)
for details, but in WWII US soldiers expressed hatred of the (Germans
| Japanese) as follows (from lowest to highest):
combat veterans (same theatre [Europe | Pacific]
combat veterans (other theatre)
trainees (in the US)
Hatred of the Japanese was uniformly higher than hatred of the
Germans, but the pattern was the same.
And on the general question of combat incentives ("Generally, from
your combat experience, what was most important to you in making you
want to keep going and do as well as you could?), "Ending the task,"
at 39%, dominated the rest. "Solidarity with the group" (14%) was
second, "Vindictiveness" was mentioned by 2%.
Mark
------------------------------
Date: Monday, 10 November 1986 08:21-EST
From: The Computer is your friend! <"NGSTL1::SHERZER%ti-eg.csnet" at RELAY.CS.NET>
To: arms-d
Re: 24 hour delay
>> From: "NGSTL1::SHERZER%ti-eg.csnet" at RELAY.CS.NET>
>> You misunderstand what I said. The tankers cannot keep THEMSELVES (not
>> to mention the bombers) in the air for 24 hours. This means that NONE
>> of the bomber force would survive.
>But they could. There is no intrinsic reason that a tanker cannot
>itself be refueled in the air.
Except that we cannot do it now. Does this mean you are saying we should
spend money buying/developing the extra tankers and bombers needed for
this policy? If so, you should say so now (and write your congresscreature).
>I have spoken to a couple of submarine officers. They have said that
>they are under orders to wait, and then try to monitor other radio
>traffic. In a pinch, they have orders to go back to the base to look.
>A submarine has many ways of getting information from the world, even
>without surfacing. Even if it does, the ocean is a big place, and
>poking an antenna above the water for 10 seconds is not likely to be
>seen.
Nuclear war counts as a pinch. This means, that all you do is wait for the
subs to go to visual range of the ports and nuc them then. The ocean is a
big place, but the area around the ports is not. Poking an antenna above
the water for 10 seconds can be deadly if the enemy knows about where you
are and has nuclear weapons.
>> the reason we have the triad is to pervent an advance in some capability
>> from destroying all our capabilities.
>In the event that the Soviets achieve a significant ASW advance, then
>I will rethink the policy. In the meantime, we should concern
>ourselves with what is true now.
Lets hope we find out about this new capability in time for you to
rethink the policy.
>The Navy has categorically stated that the present SSBN force is
>not vulnerable to current Soviet ASW. While they may elaborate on
>this in classified testimony, they will not flatly contradict it.
The Sgt. York is also the greatest thing since sliced bread (if you ask
the people in the Army with a vested interest in it). The fact is, they
are only going on an estimate of what they believe that also takes into
account their vested interest (if Soviet ASW is better, then they are out
of a job). You are betting WWIII on what you think about Soviet ASW
capability. If you were a Soviet commander, and you were told that the
American ICBM, bombers, and an optimistically large portion of the
subs could be eliminated, a first strike would look pretty good. If your
estimates were correct it would work. This adds to the instability.
------------------------------
Date: Mon, 10 Nov 1986 11:55 EST
From: LIN@XX.LCS.MIT.EDU
Subject: 24 hour delay
From: <"NGSTL1::SHERZER%ti-eg.csnet" at RELAY.CS.NET>
Except that we cannot do it now. Does this mean you are saying we should
spend money buying/developing the extra tankers and bombers needed for
this policy? If so, you should say so now (and write your
congresscreature).
Yes. I agree. I have done so.
>In the event that the Soviets achieve a significant ASW advance, then
>I will rethink the policy. In the meantime, we should concern
>ourselves with what is true now.
Lets hope we find out about this new capability in time for you to
rethink the policy.
The Soviets could have achieved a breakthrough in telepathic ways to
find submarines, and we might not know about it. There could be a
mind warfare gap!! I'm not worried. You have to make a judgment
about what is plausible.
>The Navy has categorically stated that the present SSBN force is
>not vulnerable to current Soviet ASW. While they may elaborate on
>this in classified testimony, they will not flatly contradict it.
The Sgt. York is also the greatest thing since sliced bread (if you ask
the people in the Army with a vested interest in it). The fact is, they
are only going on an estimate of what they believe that also takes into
account their vested interest (if Soviet ASW is better, then they are out
of a job).
You forget that the Navy also has control over US ASW, and you can be
very sure that if the Soviets had better ASW than we did, the Navy
would be screaming bloddy murder and demanding all kinds of money to
catch up.
You are betting WWIII on what you think about Soviet ASW capability.
Yes. You have to make a judgment about what is plausible. You also
have to take into account the advantages of what one proposes.
Waiting DOES have SOME advantages. It also has SOME disadvantages.
My view is obviously not in the majority, but I'm trying to get people
to look past their own narrow preconceptions of what is and is not
appropriate.
If you were a Soviet commander, and you were told that the
American ICBM, bombers, and an optimistically large portion of the
subs could be eliminated, a first strike would look pretty good. If your
estimates were correct it would work. This adds to the instability.
"Could be eliminated"? With what probability? What confidence would
YOU have if any analyst told you that? Would you want the Soviet
missiles and the Strategic Rocket Forces and plan a first strike on
the U.S.? What assurances would *you* need?
Or do you believe them to be crazy madmen who are far less cautious
than you are? That places your view outside the U.S. intelligence
assessments. Besides, if they are, you can't do anything about it
anyway.
------------------------------
Date: Mon 10 Nov 86 14:51:07-AST
From: Don Chiasson <CHIASSON@DREA-XX.ARPA>
Subject: Nuke equivalnce/long endurance aircraft
1. > From: Steve Walton <ametek!walton@csvax.caltech.edu>
> Subject: Meteorite as A-explosion
> ..... (I can't find the conversion from ergs to megatons in the CRC,
> even though it tells me that there are 160 square perches in an
> acre.)
'The Effects of Nuclear Weapons' (1977), p. 13 states that a one
kiloton blast is equivalent to 4.18e12 joules.
2. Several contributors have been discussing how long airplanes could
stay continuously airborne after a nuclear strike. Figures of 24
to 72 hours have been mentioned. Fuel may not be the problem: a
large number of nuclear explosions would raise massive dust
clouds. The dust particles would likely cause jet engines to
stall. Remember the case of a 747 near the volcano in Washington
and all its engines quit?
Don
------------------------------
Date: 10 Nov 1986 14:07:55-EST
From: Hank.Walker@gauss.ECE.CMU.EDU
Subject: Big R&D programs
I am tired of people out in Hawaii not paying attention. Many, and probably
most critics of the current SDI program are not against ballistic missile
defense research. What they are against is the program as currently
constituted. They feel that it is foolish to pump gigabucks that we don't
have into a program that is at best grossly premature and ill-advised. Most
of the money will be wasted on big demonstrations of obsolete technology
that tell us very little, and are intended primarily to give the illusion of
progress. Even some SDI researchers have said this. A good analogy would
be to start a $30B, 5-year program to explore the feasibility of sending a
manned space mission to Tau Ceti. We have so many years of the basics ahead
of us that it is silly to even think about attacking the problem of
interstellar travel directly.
------------------------------
Subject: Re: Military Institutions and AI
Date: Mon, 10 Nov 86 14:37:49 EST
From: Bob Munck <munck@mitre-bedford.ARPA>
toma@Sun.COM (Tom Athanasiou) writes:
> Does anyone know of institutional forces within the
> military that predispose positive receptions for
> technologies that don't really work. There's been a lot
> of talk about SDI, but I'm interested in AI per se. The
> level of hype in the commerical AI world has dropped a
> lot faster than in the military AI world. Why?
It's my belief that a great deal of the belief in AI is due to sheer
frustration with conventional ways of acquiring software. No one that I
know of considers software an easy thing to come by, but the military,
for various institutional reasons, has immensely more difficulty than
the "non-military."
Now they have been offered a MESSIAH by purveyors of AI hype: software
that works perfectly without cost and schedule overruns, slippery
accounting, forked-tongued marketeers, and bearded, blue-jeaned
programmers who aren't at their desk at 0745 hours. (A neat twist: much
of the hacker culture of jeans, beards, Chinese food, etc. originated in
the AI Lab in Tech Square.) Military types are trained to look for
alternate ways to win a battle that's going badly; of course they're
going to seize on this deus-ex-machina as a way of avoiding the troubles
of software.
Another factor: as time goes on and production AI systems continue to
fail to appear, the commercial world has less and less money available
for them. On the military side, the Reagan/SDI feeding frenzy is still
in full swing. So naturally that's where those selling AI concentrate
their efforts.
-- Bob Munck
disclaimer: none of my opinions are ever even
vaguely related to MITRE policy.
------------------------------
Date: Sat, 8 Nov 86 12:41:00 EST
From: strig@firenze.bellcore.com (lorenzo strigini)
Subject: Re: SDI is impossible
In Arms-D, V7, #47, Jim Hunt states: SDI IS IMPOSSIBLE. His technical
statements seem correct (though I am no weapons expert), and have
interesting political consequences.
His argument (as I understand it) is: SDI cannot work against weapons
engineered for survival; and weapons built for low-precision counter-
population strikes can easily be made survivable.
A consequence seems to be that one could try and develop an SDI without
destabilizing effects. That is, the specifications would be reduced to
"Make life difficult for high-precision weapons". The U.S. would tell
the USSR: "If you launch a first strike, it will kill a lot of people,
but our Strategic Defense will preserve our ability to retaliate. If
your ICBMs are meant for retaliation only, SDI will not change their
effectiveness, and you need not worry about SDI allowing us to launch a
first strike". This might improve stability by allowing the US to have
less ICBMs, longer decision times before retaliation, etc.
It would not work if: 1) the weapons that are best for defense against
precision strikes are also very good against counter-population
retaliation; or 2) even such a reduced SDI is too costly or unfeasible.
I have heard statements that "real SDI" (not the "impotent and obsolete"
version) would only protect the U.S. retaliatory force, but no statement
to the effect that it might be inherently non-offensive for
technological reasons. I would like comments from knowledgeable people.
Lorenzo Strigini
(strig@lafite.bellcore.com)
[PS to the moderator. I will not be able to read ARMS-D for
some time. Persons that wish to discuss with me and want a prompt
response should write to me personally].
------------------------------
Date: Mon, 10 Nov 86 14:39:25 est
From: drogers%farg.umich.csnet@umix.cc.umich.edu (David Rogers)
Subject: portable anti-tank weapons
On the slightly dull topic of portable anti-tank weapons, does anyone
know about the (I believe) Swedish anti-tank weapon scheduled to
replace the LAW, the standard American light anti-tank weapon? If
cheap and even partially effective (alas, the LAW was rumored to give
the occupants of newer Soviet tanks a thrill, but no real damage), it
would certainly be of use in tank-heavy Europe.
On the issue of anti-tank weapons, I am amazed by the seductiveness of
nuclear solutions (the neutron bomb) to the problem of Soviet tanks;
does anyone know the cost of the program the develop that weapon? If,
as Herb Lin claims, a billion could have developed a nice anti-tank
weapon, why was the money not spent on that instead? And indeed, once
the neutron bomb was found to be undeployable, wouldn't it have been
the obvious choice to shift the research funding to the less-glamorous
technology of portable anti-tank weapons?
David Rogers
drogers@farg.UMICH.CSNET
------------------------------
Date: Monday, 3 November 1986 08:05-EST
From: Jane Hesketh <jane%aiva.edinburgh.ac.uk at Cs.Ucl.AC.UK>
To: ARMS-D
Re: Star Wars flawed #6-of-10
Limitations of Simulation, Self-Test, and Sensor Technologies,
and their effects on the Strategic Defence Initiative
Phil Odor
Context and Problem Summary
For an electronic computer-based weapon system to function
correctly when needed, it must be
o+ completely and correctly specified;
o+ correctly built (or written) to that specification, and
the product verified against it;
o+ given access to all data needed to perform as
specified. This data represents
o+ the world-state (via sensors/communication)
o+ the local and global system state (via on-board
test equipment and communications).
This paper focuses on three sets of issues, beginning with
those arising out of the use of simulation to give
confidence in the system design. Since formal design methods
don't address the problem of correct initial specification,
the design process will include testing under simulation,
implying that the simulation itself
o+ is completely and correctly specified
o+ is correctly built
o+ has all necessary data.
We argue that simulation techniques cannot supplement or
expose a design based on false premises, and that in the
case of SDI cannot be made comprehensive even for the design
as envisaged.
The second issue concerns information about the world-
state, and in particular that derived from space-borne
sensors. We argue that current or projected technologies for
sensors aren't reliable enough to avoid potentially
catastrophic false alarms; that they impose economically
unattainable computational and communication loads; that
they are (and will remain) vulnerable to increasingly
sophisticated decoys; and that the programme calls for
time-tabled breakthroughs in sensor technology which cannot
be guaranteed.
Finally, there are issues surrounding the system's
self-knowledge. Knowledge of local state is vital for
unattended space-borne components, which should
o+ function with 100% reliability for many years, or
o+ repair themselves, possibly with remote advice, or
o+ fail passively, with correct self diagnosis.
Global system status on the other hand is needed for
decision-making during presumed attack, since the threat
hypothesis cannot be sustained when the system is
unserviceable beyond a certain limit.
We argue that Built-in-Test-Equipment (BITE) systems can
in principle never be complete. Moreover, for both good and
bad engineering and economic reasons the BITE design will
not near its theoretical potential fault coverage. The
result will be both a decrease in the effectiveness of the
system during a real attack as dormant faults are exposed,
and an increased likelihood of the system being unable to
suppress one or more of its components initiating a false
counter-attack.
Simulation
In SDI there can be no equivalent to a complete weapons
system trial: hence the reliance on simulation during
design.
You can only simulate what you can imagine can happen:
if a situation wasn't envisaged during the system
specification, only serendipity or error will produce it in
the simulation.
You can only simulate what you can afford to
reconstruct. The costs involved in a comprehensive
simulation of SDI will be prohibitive.
Simulation systems are themselves subject to error in
design or build.
Why rely on testing through simulation?
Flight tests of any aircraft prototype and test firings of
nuclear weapons both show tacit recognition that simulators
can merely represent those design parameters well understood
by designers in the previous generation of equipment, and
only an active test in a real environment gives new
information about a next-generation system. The problems
are especially acute where new technologies are employed, or
where large systems with autonomous components are being
implemented. In the case of SDI however this form of
testing is impossible.
Why is simulation not sufficient?
Other presentations will cover the impossibility of
verifying the completeness and correctness of the system
specification. The principles underpinning those discussions
extend to the effectiveness of any simulation test rigs
which are devised, since any limitation in vision and
inventiveness in formulating the original system
specification will also constrain the coverage of the
simulation. Moreover, any errors in specification will
propagate into the specification for the simulators.
Therefore only limited further confidence in the system
specification itself can be gleaned from simulations alone.
This situation is not novel: it is one reason for the flight
tests and nuclear trials.
Further questions arise from the completeness of any
simulation systems which are built. Systems which are
devised to test a design by simulating its environment are
themselves costly. The costs and complexity of a total
system test rig will undoubtably militate against its
creation. For economic and engineering complexity reasons
as much as for failures in foresight, the total system will
therefore be delivered untested.
The specification of the system will include a
failure/functionality mode analysis, which will presumably
be ratified using simulation of foreseeable failures, and
which will rely in part on the effectiveness of the system
built-in-test for information. We will return to the
correctness of information available to the system;
however, we note here that the construction of an adequate
specification of the reversionary modes of a system of such
size is an engineering task of enormous magnitude, far
greater than anything undertaken before. The scale of the
exercise again makes it very unlikely that reversionary mode
simulation and testing could be afforded.
Sensor Technology
New sensor technology will reduce, but not eliminate, data
degradation due to overload, interference and noise.
The performance limitations will be exacerbated in SDI
sensor systems by operating near technology limits; in
difficult, uncontrollable, "dirty" environments; under
stress; and with decoy systems of increasing sophistication.
Conventional techniques to reduce failures (such as sensor
duplication) would be prohibitively expensive in an SDI
setting of some thousands of sensors. Some well understood
approaches (such as averaging methods) may not be usable due
to time constraints. Functional failures of the sensors
themselves are therefore inevitable, leading to loss of
target, generation of false targets, or misrepresentation of
the speed or position of a target. There are two
ramifications:
o+ the size of SDI grows multiplicatively with numbers of
targets, without achieving 100% effectiveness;
o+ the unreliable data inputs will degrade performance in
conflict and decrease the system's safety in peace.
Doppler sensing, discrete difference detection and smear
detection all produce results which are computationally
marginal and difficult to interpret even for humans.
Artificial vision theory and engineering are not
advanced enough to guarantee object identification.
The SDI programme calls for time-tabled breakthroughs in
sensor and data handling technology which cannot be
guaranteed.
The computational demands created by the system as a
whole are on a scale which has not been demonstrated before,
even in a benign environment. No engineering models are
available for a multiple-autonomous-processor computer
system with even one thousandth of the processors needed.
The costs of the sensor system alone make its full
implementation unlikely, even given military budgets.
Lesser implementations will be correspondingly less safe and
less effective.
Hence, SDI will be neither technically satisfactory, nor
cost-effective as a complete shield against massive nuclear
attack: it will also be an unsafe implementation of a
partial defence of military targets.
These issues are dealt with more fully in "Why SDI's
Perceptual System will be unsuitable", attached.
Built-in-test systems
Information from BITE in sensor/communications-based systems
can in principle never be complete.
During the system design stage, BITE (and therefore
system state information) will be traded off in the design
specification for good engineering reasons against weight,
complexity (and therefore inherent reliability and system
integrity).
During development, cost control practices will reduce
BITE functionality: BITE necessarily appears later in the
detailed design than other aspects, and suffers most from
time and cost over-run.
Multiple subcontractors compound the problem: subsystem
BITE specifications have historically been met by testing
what is easy (crude main processor functions, RAM, ROM,
multiple system comparators, time-out monitors) and ignoring
what is hard (intersubsystem communications, sensor
functioning, actuation system functions).
BITE is relied on not only for subsystem fault-
tolerance, but also total system integrity. Poor
implementations will increase the risk of total system
malfunction, or premature weapons deployment. It is likely
that BITE will be poor.
Why will BITE be incorporated?
SDI system components will fail, as all engineered systems
fail. They will do so because of design failures, or
because the enemy has altered the operating environment, or
because of wear-out, or overstress, or damage (battle or
accidental). In other environments such failures are dealt
with by preventative or on-condition maintenance, and the
system reliability includes calculation of the effectiveness
of these procedures. One side effect of such human
intervention is increased information about the system state
as a whole, thus reducing the potential for any particular
component failure to propagate its effects through the rest
of the system.
However, the most critical SDI systems will be non-
maintainable, space based components. The system's
knowledge of its own serviceability will depend upon direct
information from BITE on failed components, and inferred
information from co-operating components. What to do under
the conditions of single or multiple failure will be a major
design problem, depending for information on BITE and good
communications.
Will BITE provide good information?
There are good reasons to believe that BITE systems will be
unsatisfactory providers of system information, for the
following reasons.
BITE information is in principle incomplete.
o+ Full weapon functionality cannot be tested without
firing the system. (Note that this is not the same
issue as that of testing during development.)
o+ The system can only infer sensor effectiveness (using
non-target information). It will be a major enemy
activity to introduce misinformation during the long
life of the system which will add to the "lie count".
o+ Only in rare cases does BITE not introduce more
equipment than in a basic system. This equipment itself
needs testing, and in the case of SDI, away from a
maintenance line.
The potential for incorporating BITE will be severely
reduced in the SDI environment. High-confidence techniques
such as "stem to stern" signal injection/actuator response
tests will be discouraged since they reduce the integrity of
the system by both reducing system up-time, and more
importantly, introduce the possibility of catastrophic error
during or after the test itself.
BITE introduces more components, and thus reduces
reliability. Again, this situation works against making
available vital information on system state.
Finally, the system will be produced by manufacturers in
collaboration with each other through the normal sub-
contractor systems. The cost constraints acting in this
period will again reduce the effectiveness of BITE, since
manufacturers are concerned to achieve their own specified
targets for BITE coverage in the simplest fashion. Since
the achievable targets cannot be specified as 100%
functionality, and since traditionally loss of information
has taken a low priority against integrity and reliability,
there is no reason to suppose that the subcontractors will
focus too hard on the difficult areas of sensor test,
actuator test and inter-subsystem (i.e. inter-manufacturer)
testing. These economic and practical issues are not about
to be overtaken by new technological breakthroughs: in
1984, a representative from one major avionic system design
firm gave this view of BITE:
"false alarm probability is higher than was expect-
ed; BIT capabilities are less than specified; BIT is
not totally independent of operators interpretation
.."
What is expected of BITE, and what are the results of its
limitations?
A failed component will rely on its BITE to do three things:
o+ to inform the system as a whole;
o+ to adopt an appropriate reversionary mode if possible;
o+ to fail passively if no reversions are available.
SDI sources now claim that it will be sufficient to create a
system with less than complete effectiveness for SDI to be
counted as an overall success. However, the implications
for world security of designing systems in which there are
potential active failure modes is clear. Relying on current
levels of BITE technology to inform on failure modes will
increase the chances of such a failure.
As far as the system as a whole is concerned,
information from a component can be present (and true or
false) or absent (e.g. if outgoing communications fail).
Even if BITE is used, therefore, the system's resolution of
its own state is not a trivial task: it certainly involves a
ternary logic system which includes "liars". The less
effective BITE and the communication channels are, the more
potential liars in the system, and the more chance of
misinterpretation. It is not clear that a reliable
algorithm for handling such a system is feasible.
Information about the author
J. P. Odor BSc(Hons) Computing, Senior Research Fellow in
Computing and Education, Education Department, University of
Edinburgh.
Director of Communication Aids for Language and Learning
Project.
Research Interests:
Computer Modelling of student learning, automated item
banking techniques for skill and academic assessment, uses
of microelectronic systems for handicapped learners,
especially the production of intelligent communication aids
and smart wheelchairs.
Service Career:
RAF 1962-1973, Navigational Instruments. Member of Central
Servicing Development Establishment 1969-1973, permanently
detached to BAC Warton during Tornado development.
------------------------------
End of Arms-Discussion Digest
*****************************