Slashdot Mirror
The Big Technical Mistakes of History
An anonymous reader tips a PC Authority review of some of the biggest technical goofs of all time. "As any computer programmer will tell you, some of the most confusing and complex issues can stem from the simplest of errors. This article looking back at history's big technical mistakes includes some interesting trivia, such as NASA's failure to convert measurements to metric, resulting in the Mars Climate Orbiter being torn apart by the Martian atmosphere. Then there is the infamous Intel Pentium floating point fiasco, which cost the company $450m in direct costs, a battering on the world's stock exchanges, and a huge black mark on its reputation. Also on the list is Iridium, the global satellite phone network that promised to make phones work anywhere on the planet, but required 77 satellites to be launched into space."
Rim shot...!
No sig today...
There was no technical flaw in Iridium. It was stated what it would do. It did it. Someone screwed up the business plan, but there was no technical mistake. They knew it took 77 satellites for what they wanted. And they launched them all and they worked flawlessly. Now, if only they had sales to match the business plan, they'd be billionaires. But again, unrelated to any technical issue.
Learn to love Alaska
Don't forget the Therac-25
Poor software design and development led to radiation overdoses for 6 patients being treated for cancer, with 3 dying as a direct result.
Sadly, mistakes still keep on happening.
They forgot the cd protection cracked with a black marker...
http://www.zeropaid.com/news/1069/black_marker_cracks_cd_protection/
European Linux user, living in Antwerp
When I saw the title, I immediately imagined the Maginot line. Thousands more examples could come to mind.
Could somebody please explain to the author of the articles that Technology is more than computers/gadjets and older than 10 years? It is an epic history that goes along with mankind.
What, no Capacitor Plague? http://en.wikipedia.org/wiki/Capacitor_plague
The technical error here was that there was no test on the real thing. The company that made a part of the telescope had only a separate testbed that was made to specifications. Alas, these specifications were exactly one inch misunderstood, so the result was a part that was incredibly accurately one inch out of position.
Nae king! Nae laird! Nae yurrupiean pressedent! We willna be fooled again!
Seriously, we have got to stop with the hyperbole before our children don't know the difference between a War on Drugs and a War in Iraq.
We we say of all time, I think of things like lead plumbing in Rome, or the suspension bridge that got tore apart by a mere breeze.
http://en.wikipedia.org/wiki/Lead_poisoning#History
http://video.google.com/videoplay?docid=-3932185696812733207#
Always going forward, 'cause we can't find reverse.
The problem Intel had with the FDIV bug was one of PR. The Pentium range was the first CPU family to be directly marketed to the general public in a big way.
While anyone with knowledge of the chip design and production processes understood that such bugs are not particularly uncommon (many much simpler chips have well documented errata and workarounds for unintentional behaviour, like the 286's "gate A20" bug that actually turned out to be useful) the general public and the popular press had no such understanding so were very surprised - they assumed that all CPUs were (or should be) completely 100% perfect and therefore taking issue with what they saw as being sold defective goods.
Before the first generation Pentium FDIV issue, such relatively minor problems were dealt with by the error, including any extra side-effects and possible workarounds, being documented, those errata being sent to the chip makers customers and relevant software developers, and things would get patched up without the general public ever being aware there was an issue in the first place aside perhaps from a small number of users who by shear chance were noticeably affected by the one-in-a-few-billion problem before their software was patched (those people would be given replacement chips and/or other recompense). A costly replacement program simply wouldn't have been needed in this case.
Speaking of technical flaws and Lisa... You could plop the boot drive into the Dumpster, and it would format it. The tech savvy devs who designed the "drag-to-trash = format" function never imagined that users would be stupid enough to do something like that! Little did they know about how giving someone a mouse transforms them from someone who can use a line based editor to set up printer drivers and networking into the horror that is a modern user.
Where's Microsoft Bob? Novell Groupwise? Lotus Word Pro? Lantastic?
There's that, and there's also the whole "the world is flat" and "disease is caused by imbalances in the four humours of the body" ideas. The article's examples seem pretty trivial in comparison.
... and then they built the supercollider.
The virus is thought to have been developed in 1986 by two brothers in Pakistan named Basit and Amjad Farooq Alvi, who were looking to protect some medical software they had written from disc copying. They had found some suitable code on an internet bulletin board site and adapted it so that if someone used the software then the malware would be installed.
I'm guessing "Iain Thomson" is not a day over 25, not very versed on the history of the Internet, and too busy to look up the meaning of "BBS". Am I right?
This post contains no rudeness or derision of any kind. All arguments are friendly. Terms and exclusions may apply.
I still have one of the Pentium 90 chips with the math flaw. The bidding starts at $1.
When I tried to work it out it came out as $449.9999867' million.
Confucius say, "Find worm in apple - bad. Find half a worm - worse."
Intel's 8086 CPU, Intel's first 16-bit processor, was possibly much worse than any of those mentioned because it affected all of us. Intel chose to continue the quirkiness of the 8008 rather than abandon it.
... and was therefore fairly awkward (and remained so until the 80386)."
Just before the time of the introduction of the 8086 I knew a chief of technology of a high-tech company who was waiting for the 8086 as though it were a combination of Christmas, his birthday, and the birth of his child. He would start every conversation by telling everyone Intel's release date for the 8086.
The day of its release, he was miserably unhappy. Intel chose to continue an architecture that made assembly language programming and debugging of high-level languages more difficult.
Wikipedia says about the 8086: "Marketed as source compatible, the 8086 was designed so that assembly language for the 8008, 8080, or 8085 could be automatically converted into equivalent (sub-optimal) 8086 source code, with little or no hand-editing. The programming model and instruction set was (loosely) based on the 8080 in order to make this possible. However, the 8086 design was expanded to support full 16-bit processing, instead of the fairly basic 16-bit capabilities of the 8080/8085."
The problem was that the quirkiness has been extended to the 32-bit processors of today. The Wikipedia article says, "The legacy of the 8086 is enduring in the basic instruction set of today's personal computers and servers..."
And, "Programming over 64 KB boundaries involved adjusting segment registers
Everyone on the planet who used or were affected by computers then suffered because the debugging was much more complicated than if Intel had chosen to make the operation of the 8086 simpler.
"Such relatively simple and low-power 8086-compatible processors in CMOS are still used in embedded systems."
A man I worked for many years ago, one of my engineering mentors, told me about a mistake made during World War Two, where a large number of very large castings were discarded because the specification called for a much smaller tolerance on the location of an exhaust port than was actually necessary. As I recall, the spec allowed it to be 1/4" away from its nominal location, but it actually was connected to a flexible hose and it could have been a couple inches off in any direction without causing any problem. This mistake wasn't discovered until several millions of dollars worth of tank bodies had been scrapped and melted down unnecessarily.
-jcr
The only title of honor that a tyrant can grant is "Enemy of the State."
Though wasn't the issue in case of Pentium FDIV bug specifically that Intel didn't publish the errata or...any other information after Intel researchers discovered the error? It took one independent one, to whom Intel didn't even respond initially...
One that hath name thou can not otter
Yeah, I would immediately classify any error that caused deaths to be more important.
Another interesting case was the Patriot Missile failure. The system clock counted in 1/10th second increments. However, it added 0.1 to a floating point number. Unfortunately, 0.1 in binary is a repeating number, similar to 1/3rd in binary being 0.333333333...
So, ten times every second the time drifted just the tiniest bit. The missile that missed had been running for days, so its clock was one third of a second off, and a Scud travels a long way during that time.
Let that be a lesson to all of you: use an integer counter, and divide by 10 to get the time in seconds.
Write your representatives! Repeal the 2nd Law of Thermodynamics!
To quote Wikipedia:
The metric/imperial mix-up that destroyed the craft was caused by a human error in the software development, back on Earth. The thrusters on the spacecraft, which were intended to control its rate of rotation, were controlled by a computer that underestimated the effect of the thrusters by a factor of 4.45. This is the ratio between a pound force–the standard unit of force in the imperial system–and a newton, the standard unit in the metric system. The software was working in pounds force, while the spacecraft expected figures in newtons; 1 pound force equals approximately 4.45 newtons.
The software had been adapted from use on the earlier Mars Climate Orbiter, and was not adequately tested before launch. The navigation data provided by this software was also not cross-checked while in flight. The Mars Climate Orbiter thus drifted off course during its voyage and entered a much lower orbit than planned, and was destroyed by atmospheric friction.
Source: http://en.wikipedia.org/wiki/Mars_Climate_Orbiter
One-in-a-few-billion problem ?
At that time, I was programming a network game about trucks, and when when replaying a demo on the network, the players desynchronized after a few minutes.
I spent a lot of time looking into the logs, and discovered that there was a floating point error that desynchronized the trucks.
I still believe that the FDIV bug was much more frequent than publicized, and it had more impact than what Intel originally described.
Intel released a software patch to Watcom C++ library, but the patch was terrible, with the FDIV replaced with a lot of instructions just to detect the cases where the bug might appear, and use shiftings instead of FDIV.
I think that the bug was much publicized because it was the beginning of Internet, where a lot of new information went unfiltered, and Intel completely missed their communication on this bug discovered by Thomas Nicely.
Here is the whole story behind this bug:
http://www.trnicely.net/pentbug/pentbug.html