From:         "peter (p.j.) ashwood-smith" <>
Organization: Bell-Northern Research, Ottawa, Canada
Date:         01 Jun 95 05:00:52 
References:   1
Next article
View raw article
  or MIME structure

In article <airliners.1995.667@ohare.Chicago.COM>, (Thornton Shepherd) writes:
|> (Jean-Francois Bosc) wrote:
|> >In article <airliners.1995.627@ohare.Chicago.COM>, (George Mantis) writes:
|> >
|> >> Airbus definitely gives their FBW systems more "authority" than Boeing does
|> >> with theirs. The result: an awfully high number of Airbus crashes, due to
|> >> pilots that are either unsure of the autopilot's reaction, or the pilots
|> >> "fighting" the inputs of the autopilot. In short, I'd have to say they
|> >> definitely have it wrong.
|> >
|> >Always the same kind of assertion. Some STATISTICS, please !
|> >
|> >JF
|> >
|> >PS : Since this post will probably be censored like the previous ones,
|> >I think I can comment about the amount of unsubstantiated allegations
|> >against Airbus that ARE NOT censored on this group. I personally am just
|> >requesting FACTS, ie STATISTICS. Any honesty around there ?

   First of all I doubt anybody is censoring anything here. The Internet
is notoriously unreliable as are all computer based systems ;) besides
I'm a Canadian and have no special interest in Boeing.

   Secondly there are not really enough statistics yet to add up to much.
What we have here are a large number of people with gut feelings that
relying on a computer to keep you alive instead of a person is not
necessarily the best way to go.

   Since computers were invented the general population and much of the
scientific community at large have been swept up in the power of what a
computer can do. It is only in the last 5 or so years that many people
are starting to question just how much actual "progress" we are making by
sticking a CPU into everything. The accountants can usually easily quantify
the cost savings (and weight savings) of using a computer where a
mechanical system once existed but the true cost must include the other
side of the equation which includes the cost of making a system suddenly
take on a non linear relationship between failure and the cost of that

   It is against this backdrop of "lets put a computer everywhere" that
many of us are starting to get very uneasy. Those of us that program for
a living know how nearly impossible it is to make a computer do what you
intend it to do and are constantly surprised by how it reacts to inputs
we did not anticipate.  The Airbus is just one of many systems that
personally worry me. I worry equally about large banking transactions,
hospital equipment, nuclear reactors etc. but since I have a great fondness
for flight I take the Airbus concept of a computer limited flight envelop
as a very dangerous one. All it takes is a bit error somewhere and that
flight envelop changes by a power of 2! It is this non-linear reaction to
a failure that makes the safety of such systems hard to determine and
which makes those of us that have seen bits get twiddled really scared.

   The sidestick is another kettle of fish altogether. The problem here
is that Airbus has chosen to remove the tactile feedback that has until
this time been available to the pilot. This tactile feedback which used
to tell you simultaneously what the plane/autopilot and co-pilot were
doing has been replaced by a computer screen which you have to be looking
at to figure out what is going on. Previously the pilot/copilot and
autopilot could all talk to each other about what the plane was doing
without ever exchanging a glance or word. Just put your hands on the
controls and you know how much yaw/pitch/bank and power are currently
applied and you can feel the rate at which they are changing. Control can
be easily passed from auto pilot to pilot to co-pilot with little problems,
since you know what the starting point is. All this can be done while
looking out the window for a place to land or while worrying about some
other problem. Tactile feedback is a very powerful thing and human touch
is a remarkably well developed sense that Airbus has chosen not to
incorporate into its interfaces.  I believe this is a mistake.


   Peter J. Ashwood-Smith