Slashdot Mirror
Software Glitch Caused Crash of Airbus A400M Military Transport Aircraft
An anonymous reader writes: A software glitch caused the crash of an Airbus A400M military transport aircraft, claims German newspaper Der Spiegel (Google translation). The accident, which happened in Seville on the vehicle's first production test flight on 9 May, killed four crew members. Airbus is investigating the system controlling the aircraft's engines. The early suspicions are that it was an installation problem, rather than a design problem.
The accident, which happened in Seville on the vehicle's first production test flight on 9 May,
They WERE testing the plane. cant know about the bugs until the real world tests
have you seen my sig? there are many others like it but none that are the same
ONE_IN_FAILURE_RATE = 50000000; //Ted: reduce by 10 every time management claims they need to increase reliabilty
//TODO: Ted, MAKE SURE YOU REMOVE THIS BEFORE SOMEONE ACTUALLY FLIES.
if(left_engine_running && (rand()%FAILURE_RATE == 0))
When my roommate had Comcast for cable and Internet, we always knew we would have trouble whenever a Comcast truck drove through the neighborhood. One day a technician installed something into the box out front. That killed our Internet service. Took a month to convince Comcast that the problem was outside in the box and not inside between the chair and keyboard. When they finally sent a technician out, he discovered that the last technician installed the bypass filter backwards.
There's nothing but wild speculation what happened in that crash, but most of it focuses on the severe weather the plane was passing through. And this is the second time in two days you posted the same comment anonymously; what's your point?
The wording is somewhat ambiguous. It was the first flight *of that specific aircraft*, not the first flight of an aircraft of that kind.
Isn't an "installation problem" by definition a design problem?
Aren't we passed the days of process not being part of design?
Depends. Did whoever follow the process? The design/process should make it easy, but the world is always inventing better idiots.
Seriously.
If your shit can be installed wrong and lives depend on it being installed correctly, it's designed wrong.
We aren't inventing better idiots. Designers continually fail by underestimating the ingenuity of idiots.
Don't think anyone would be stupid enough to plug it in backwards and pound it in with a hammer? You've just underestimated the idiot.
Designers and engineers are generally too logical to see all the failure paths that someone could take because they don't make the assumption that it will be installed upside down or with a hammer. It's the simple illogical design processes that prevent these extreme events, such as making the part fit in only one direction (sizing the part such that it cant physically fit or be connected) or such as USB C making it fit either direction. These design strategies appear to make the design illogical or more expensive for no logical reason so they aren't implemented by the logical engineer/designer. But these illogical processes are how you prevent the ingenuity of idiots from getting the better of you.
The Bell Boeing V-22 Osprey crashed 4 times during testing killing 30 crew members. The previous Airbus crash in testing was in 1994, if you want to go that far back there was a fatal Antonov An-70 crash in 2001, also due to engine problems. New aircraft sometimes crash as the bugs are worked out, the 787 was just lucky that none of the incidents were fatal.
All modern planes except light GA aircraft have engines have fully computer-controlled engines, it's called FADEC and it's what makes them efficient, reliable and much safer (in general). Sometimes these have bugs, particularly on new engine designs.
Some thing appear to have been lost in translation.
According to most other English language sources, apparently this A400 had a new software release that enabled it to control the fuel tank trim during some new types of maneuvers. It appears that some bug in this software triggered a situation where fuel was actually cut-off from the engines or perhaps the engines shut-off leading to a temporary engine stall (which proved to be unrecoverable). It's not clear exactly what happened yet, but I think they are close to ruling out a defect in the installed ECU (electronic control unit) itself, but not the software running on it.
Scott Adams' Falacy #24: IGNORING ALL ANECDOTAL EVIDENCE
Example: I always get hives immediately after eating strawberries. But without a scientifically controlled experiment, it’s not reliable data. So I continue to eat strawberries every day, since I can’t tell if they cause hives.
Your assumptions are somewhat naive - +1 for the slashdot-populist-anger though.
Aircraft are not plug-and-play systems, like your home computer or tablet. They consist of miles of wiring and cabling, as well as hundreds of sensors, on board electronic computers and mechanical fail-safes from many different suppliers all over the world. For these reasons (as well as technological and environmental issues) not every single computer from every single company talks on the same hardware interface, or uses the same protocol. Sometimes, signals between devices are best suited to be analog, and sometimes it's digital. Sometimes the bus is serial, and sometimes it's Ethernet-based (e.g. AFDX).
Any modern airframer (like Airbus or Boeing) will have drawing and schematics on where a particular device is to be installed and how each pin of each device's connector (including power, inbound and output communication buses) should be wired to the rest of avionics. But, I would assume, that in this case someone made a mistake in the final assembly and someone else in quality assurance dropped the ball by not catching it.
Designing and integrating everything into a modern aircraft is one of the most complex tasks an engineer can experience. It's super complicated - But that does not mean that is "designed wrong".
Sometimes, people just make mistakes.
"Airbus Defence and Space has today (Tuesday 19 May) sent an Alert Operator Transmission (AOT) to all operators of the A400M informing them about specific checks to be performed on the fleet.
To avoid potential risks in any future flights, Airbus Defence and Space has informed the operators about necessary actions to take. In addition, these results have immediately been shared with the official investigation team."
What exactly was the contents of the (AOT). What specific checks were required. Is this related to the 248 day rollover bug in the Boeing 787 generator control units (GCUs)?
Statement regarding Alert Operator Transmission (AOT) to A400M operators
Since it was the first flight, the EULA popped up, and the crew made the mistake of hitting "decline" instead of "accept"?
Ironically, no FADEC software is French. However, I presume you masturbate to Boeing planes without until now knowing that they're designed exclusively on French software (CATIA by Dassault Systems).
Exactly who is responsible for this kind of software bug?
The coder who wrote the code?
The functional spec writer?
The QA tester who didn't catch it?
The test scenario scripter?
The manager who oversaw the development process?
The QA manager?
The stakeholder who OK'd the move to production?
The project manager who co-ordinated the project?
The CTO of the company who funded the effort?
Or should they all be criminally liable, thus diluting the responsibility of any one person so that no one person is actually liable?
I sense a stone-thrower in a glass house here...
Sent from my ENIAC
According to an article in Spiegel Online three of the engines shut down during takeoff.
There were also claims that much of the software was written by underpaid inexperienced developers and there was high turn over due to a high pressure environment.
Why is Snark Required?