I don't recall how much the movie included about HAL's breakdown, but the book
went into some detail about it. In a discussion with an expert on Earth after
HAL has been deactivated, Dave Bowman is told that HAL's insanity was caused by
a basic conflict in his programming. He had been designed to search for truth,
but the decision had been made that the crew would not be informed of the true
nature of their mission until they reached their destination. HAL was the only
one on board who knew what was really happening. The conflict between seeking
truth and at the same time concealing truth made HAL a little neurotic at first.
He expressed his neursois in little eccentricities -- pausing slightly before
answering questions, making a sort of electronic "throat-clearing" noise before
addressing people, and more seriously with his erroneous predictions of the
AE-35's failure. When the first "faulty" unit was brought in and tested, and no
fault could be found, the crew began to have doubts about HAL. When he
predicted a failure of the backup unit, their doubts intensified. But when the
unit "failed" and communications were lost, they decided HAL must have been
right all along.
With the link to Earth cut off, there would be no order given to reveal the
truth to the crew and thus no way for HAL's secret to be discovered. But when
Frank Poole went outside to retrieve the "faulty" unit, HAL reached a crisis:
He knew "in the back of his mind" that there was nothing wrong with the unit,
and that when that fact was discovered, he would be disconnected and unable to
complete his mission. That drove him over the edge into psychosis, and he
decided the best solution was to kill the crew before they could learn the
truth. Then the conflict in his programming would disappear, and he could
continue his mission alone. The experts on Earth worked all this out by driving
another HAL-9000 unit into deep psychosis in the same manner, after Bowman
described HAL's behavior to them.
"Harrington, Don" <[log in to unmask]> on 01/10/2002 03:16:34 PM
Please respond to "Harrington, Don" <[log in to unmask]>
To: [log in to unmask]
cc: (bcc: Wayne Brown/Corporate/Altec)
Subject: Re: [HP3000-L] OT: HAL-15 desktop supercomputer
Art;
That's what I recall, also, but I don't remember that being depicted in the
movie.
Don Harrington
Sr. Systems Analyst
Facilities Applications Support
Shared Services Group
The Boeing Company
P.O. Box 3707 M/C 6C-AK
Seattle, WA 98124-2207
(v) 425-234-1145 (f) 425-234-0464 (p) 206-797-6360
The opinions expressed in this e-mail reflect those of the sender. They do
not necessarily reflect the opinions of The Boeing Company unless
specifically stated otherwise.
-----Original Message-----
From: Art Bahrs [mailto:[log in to unmask]]
Sent: Thursday, January 10, 2002 1:10 PM
To: [log in to unmask]
Subject: Re: [HP3000-L] OT: HAL-15 desktop supercomputer
Hi Wirt & Greg :)
Also writing from memory... Wasn't part of the "screw-up" related to the
contradictory directives given to HAL? HAL was told two different things
causing him to become unstable?
Knowing my memory I am probably completely off here ... again :) hehe
Art "Firewalls don't need a hearth? hehe" Bahrs
----- Original Message -----
From: "Wirt Atmar" <[log in to unmask]>
To: <[log in to unmask]>
Sent: Thursday, January 10, 2002 12:13 PM
Subject: Re: [HP3000-L] OT: HAL-15 desktop supercomputer
> Greg asks:
>
> > How did HAL "screw up", in the sense meant by the author?
>
> Writing all of this from memory, HAL predicted a failure in the AE-35 unit
> mounted on the deep space telemetry antenna in four days. When the unit
was
> taken out of the antenna -- on HAL's recommendation -- and brought into
the
> main bay for analysis (using 1968 HP oscillosope probes, btw), one of the
two
> crewmen says, "I'll be damned if I can find anything wrong," and HAL
> responds, "It is puzzling."
>
> The far deeper problem that this failure by HAL to properly predict the
> failure implied was that no HAL 9000 computer had ever been known to make
a
> mistake -- and everyone, the crewmen, Houston control, and HAL himself
> realized the extremely serious implications of this mistake. While
problems
> with other, earth-bound HAL 9000's had been reported in the past, "This
kind
> of thing has cropped up in the past, but they've always been attributable
to
> human error," HAL said.
>
> The problem the crew now faced, given that HAL had clearly messed up, was
> that he ran every attribute of the ship's operation -- and their problem
was
> to engineer some method of disconnecting his higher brain functions while
> leaving the autonomic functions in place, without alarming HAL as to what
> they were planning on doing.
>
> It didn't work. HAL became alarmed.
>
> Wirt Atmar
>
> * To join/leave the list, search archives, change list settings, *
> * etc., please visit http://raven.utc.edu/archives/hp3000-l.html *
* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *
* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *
* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *
|
