The Risks Digest

The RISKS Digest

Forum on Risks to the Public in Computers and Related Systems

ACM Committee on Computers and Public Policy, Peter G. Neumann, moderator

Volume 4 Issue 80

Tuesday, 5 May 1987

Contents

o Computer Risks at the Department of Transportation
PGN
o Computerized advertising network used to fence hot circuits
PGN
o EPROMS and "Wimpy" Energy Physics
Patrick Powell
o Re: Wheels up (Richard M. Geiger, Jerry Hollombe>
o Liability for software "unless you buy our method"
John Gilmore
o Info on RISKS (comp.risks)

Computer Risks at the Department of Transportation

Peter G. Neumann <NEUMANN@CSL.SRI.COM>
Fri 1 May 87 17:02:08-PDT
In an editorial on page A.14 of the San Francisco Examiner and Chronicle on 
26 April 1987 were some comments on air passenger complaints having doubled 
in March 1987 compared with March 1986.  ``But the airlines weren't the only 
ones falling short in service.  The Department of Transportation ...  usually 
issues its complaint report monthly, but the one for March was the first since 
December.  The delay was blamed on computer malfunction.  Could it be that the
Transportation Department has the same computer system as the airlines?''
[Hey, I'm just quoting.  No responses please on whether the last sentence 
was facetious or whether the editorial writer is stupid.  PGN]

The SF Chronicle on the following day had an article noting that air
traffic controller errors at Chicago's O'Hare Airport increased 65% from
1985 to 1986, and nearly led to major disasters on several occasions.
[From a newly released congressional report.]

The SF Chronicle on 28 April quoted National Transportation Safety Board
Chairman Jim Burnett, that a forced reduction in the number of flights is
a necessary short term step to offset the recent rise in air traffic
controller errors and near-collisions.  FAA chief Donald Engen defended
the system as safe and said that Burnett does not understand it.  ``I
don't believe we should adopt a policy of restricting air commerce in this
country", he said.


Computerized advertising network used to fence hot circuits

Peter G. Neumann <NEUMANN@CSL.SRI.COM>
Tue 5 May 87 09:54:34-PDT
Richard Gaudet and William Gorgizian are accused of taking $250,000
worth of integrated circuits from a San Jose electronics company,
setting up a phony parts-supply company, advertising through a
nationwide computerized network, and distributing around the country
(via UPS -- see next message).  Along with $200,000 in computers and
burglary tools, authorities also confiscated a book entitled "The
Perfect Crime and How to Commit It".


EPROMS and "Wimpy" Energy Physics

Patrick Powell <papowell@umn-cs.arpa>
Sun, 3 May 87 19:52:14 CDT
You better believe that there is a problem shipping things via airfreight.

First,  let me give you a glimpse into semiconductors, especially the
(old) EPROM technology.  (If you know this stuff,  and disagree with
my explanation,  quit quibbling: you know what I mean).  The way that
an EPROM stores information is by "trapping" a bunch of charges in
an "insulated" region;  you can do the same thing by statically charging
up some material,  and not letting the charge drain off.  This trapped
charge can be used to modify the characteristics of a transistor:
if the charge is not there,  the transistor will turn on (off?) when asked to;
otherwise it looks like a dead duck,  and will not turn on or off.  Let the
"good" transistors represent a 1,  the bad (trapped charge) a 0.
VOILA!  Programmable Memory... just find a way to get the charge there.
Well,  that can be done by placing a fairly high voltage across the
transistor,  which causes a strong electric field,  which will BLAST
those little charges into place.  Hopefully. Sometimes you have to do
this several times,  i.e.- you have a programming cycle.

Well,  all  this depends on the charges staying there.  Do you know what
happens when a charged particle rips through a solid?  It leaves a little
chain of ionized atoms in its path;  luckily this only lasts a short time.
Zap!  there is a conducting path, and away some of those little devils race,
hither and yon.  And you lost some charge.  Do this often enough,  and
PRESTO.  No more stored information.

Now at 20 Km up there (60,000 feet: 12 miles),  you would be amazed at the
numbers of highly energetic particles.  Of course they will get "absorbed"
by the atmosphere,  but that is a Loooong ways down.  One of the reasons why
military equipment is "Radiation Hardened".

By the way, it isn't just High Energy Particles.  One of the
interesting things is that UltraViolet light puts out enough OOMPH
(highly technical physics term) to cause the charges to start leaking,
and is how you can erase a EPROM.  Well, ordinary light will also
effect semiconductors, in a similar manner.  In fact, enough light,
and a "nonconducting" transistor will start conducting.  Luckily all
those transistors are hidden away inside little opaque packages,
except for (are you ready?)  EPROMS!  which need a clear window so
they can be erased.

So here we are, with a new system, on public display.  The program was
in EPROM, and was on a board.  "Lets open the cabinet, and show people
a running system!"  This led to the main board, with it's EPROM, being
exposed to the public gaze.  And to their cameras.  Flash Cameras.
With BIG Pulse Zenon Bulbs.  10 Microsecond flash, 4 times light of
sun blah blah.  Flash! Flash!

...Parity Error!  EPROM Parity Fault!  Reset and Restart....

From the comments I heard,  this just about drove a couple of people nuts.

By the way, if you want to see if this works, try getting one of those
"singing cards" with the IC on it.  Some of them have been potted in
CLEAR jel, and you can actually see the chip.  Get the thing singing
away, and then shine a light on it.  A strobe light works best, and
you can actually hear the effect.

Patrick ("Hardware?  If it was easy to build we'd call it Software!") Powell


Re: Wheels up (RISKS 4.79)

Richard M. Geiger <prls!mips!rmg@Sun.COM>
Mon, 4 May 87 20:19:59 PDT
I once saw a Cessna light plane with a prop which had severely curled-back
blades. We asked the F.B.O. (employee of the rental company which owned it)
what had happened. We were told that the plane was equipped with
(overly-sensistive!) automatic landing gear retraction. It had hit a bump
while taxiing, and bounced; the mechanism decided that the plane had taken
off, and raised the gear. Didn't do the engine much good.

Rich Geiger    {decvax,ucbvax,ihnp4}!decwrl!mips!rmg
MIPS Computer Systems, 930 E. Arques, Sunnyvale, CA 94086
(408) 720-1700 x308 [Day Job]  (408) 739-7911 [home]


re: Wheels up

The Polymath <ames!hollombe@ttidca> <Jerry Hollombe>
Tue, 5 May 87 13:21:18 PDT
As a pilot and aircraft mechanic I can tell you this is not a realistic
simulation.  All aircraft with retractable gear have a safety switch
(often called a "squat switch") that senses when any weight is on the
landing gear and interrupts power to the retraction mechanism in that
condition.  Barring electrical/mechanical failure the gear will not
retract while sitting on the ground.

A common but unsafe practice is to flip the gear control to "up" while
taxiing and allow the gear to automatically retract as the plane lifts off.

The Polymath (aka: Jerry Hollombe, hollombe@TTI.COM)
Citicorp(+)TTI  3100 Ocean Park Blvd.   (213) 450-9111, x2483
Santa Monica, CA  90405 {csun|philabs|psivax|trwrb}!ttidca!hollombe


Liability for software "unless you buy our method"

John Gilmore <hoptoad.UUCP!gnu@cgl.ucsf.edu>
Mon, 4 May 87 01:22:46 PDT
#define slime people /* For the squeamish */ 
Oho!  The slime who are in business to tell you how to take risks
("pay us money to assume them") and have coerced the government(s)
into making it illegal to do otherwise, are now joined by the slime
who are in business to tell you how to build software ("pay us money
to use our formal design software") and are now attempting to get
government guns to enforce their methods too.

                    [Somewhat overstated, but certainly a risk!  PGN]

Please report problems with the web pages to the maintainer

Top