Re: Fremtidens fly blir pilotløse
Tror ikke piloter har noe å frykte før de setter seg i et førerløst fly - men la oss ha ingeniørene ombord for prøveflygingene (-:
Forresten, ubemannede fartøyer til Mars, det minner meg på en morosom historie fra forrige århundrede (de litt sarkastiske kommentarene er fra en som har kommentert litt på websiden). Det viser at siden datamaskiner ikke programmerer datamaskiner, er "datafeil" også som regel "menneskelige feil". Men i motsetning til flygere, som i de fleste tilfeller selv er til stede for å rette opp sine egne og andre sine menneskelige feil, er dataprogrammereren ikke til stede for å rette opp feilen når en datakommando eller et datasystem slår feil:
---
NASA Says Human Error Caused Loss Of Mars Craft
The following article appeared on Yahoo! News on September 30, 1999. The [parenthetical remarks from one of the SCE faculty about these passages] are, of course, not part of the original article. Luckily, software catastrophes are still mostly (but not entirely) about the loss of money rather than the loss of life. What's your view? How much longer will it be before software systems that fly planes, control missile launches, and monitor nuclear power plants are featured in such stories? Can anything be done to make software design errors much less likely? Or are large software systems inherently so complex that we should never expect them to work correctly?
Yahoo! News
Thursday September 30, 1999 6:49 PM ET
By Michael Miller
PASADENA, Calif. (Reuters) - Human error stemming from space engineers ["space engineers"?!?] using two sets of measurements -- one utilizing miles and the other kilometers -- caused the loss of the Mars Climate Orbiter spacecraft last week, NASA said Thursday. [You're kidding, right?!?]
The teams, located at the National Aeronautics and Space Administration's Jet Propulsion Laboratory in Pasadena and at Lockheed Martin Astronautics in Colorado, complicated matters further by failing to realize the error, the agency said in a statement. [If they had realized there was an error, presumably they would have fixed it. So, of course, it isn't making the error (surely human errors are inevitable), but failing to realize it, that is the real problem.]
The $125 million orbiter, intended to serve as the first interplanetary weather satellite, is believed to have broken up when it hit the Martian atmosphere last week after an approach that was too near the surface.
``People sometimes make errors,'' said Edward Weiler, NASA's associate administrator for space science. ``The problem here was not the error, it was the failure of NASA's systems engineering, and the checks and balances in our processes to detect the error. That's why we lost the spacecraft.'' [Right!]
An investigation into the loss of the craft was launched immediately after the spacecraft was lost. A peer review board Thursday announced its preliminary findings.
The review board said that in making a key change to the spacecraft's trajectory one team used the English, or avoirdupois, system of measuring, which utilizes miles, yards, feet and inches as well as pounds and ounces, while the other was using metric kilometers, meters, kilograms and grams. [Reporter: "Oops, there goes $125 million... Too bad the 'space engineers' didn't write interface contract specifications so each team knew what the other's software was supposed to do." NASA spokesperson: "But it would have cost thousands of dollars more to hire software engineers, who certainly would know that you should do this, rather than having physicists and mechanical engineers design and write the software." -- OR -- "But it would have cost thousands of dollars more to hire software engineers who had paid attention to their mathematics and physics courses in school." Wonder which one it was...]
In a statement, the Jet Propulsion Laboratory said, ``This information was critical to the maneuvers required to place the spacecraft in the proper Mars orbit.'' [Really?!?

]
There are 1.6 kilometers to a mile and 1.1 yards in a meter, while there are 2.2 pounds in a kilogram.
At the time the spacecraft was lost, Mars Climate Orbiter Project Manager Richard Cook said scientists had expected that the orbiter would approach Mars at an altitude of between 87 and 93 miles when it fact it came in at 37 miles above the surface of the planet. He said the minimum survival altitude was 53 miles.
Jet Propulsion Laboratory Director Edward Stone said, ``Our inability to recognize and correct this simple error has had major implications. We have underway a thorough investigation to understand this issue.'' [Surprise! This is how software is: one "simple error" has "major implications". It's pretty easy to understand the source of the problem, given the above the explanation of what went wrong: if you don't specify interface contracts but instead rely on wishful naming, the components probably aren't going to work properly when you put them together.]
In addition to the peer review board composed of Jet Propulsion Laboratory scientists, a second review board that includes outsiders also is looking into the cause of the loss, and an independent NASA review board is to be formed shortly.
The primary mission of the orbiter had been to monitor the Red Planet's atmosphere, surface and polar caps for one Martian year, or 687 days.
The craft also was intended as a vital link in the Mars Polar Lander mission. That craft is due to land on Mars on Dec. 3 and the climate orbiter would have acted as a relay station between the lander and scientists on Earth.
Cook said the loss of the climate orbiter would complicate the lander mission, but contingency plans already were in place for the lander to transmit data directly to Earth through the Deep Space Network and via the Mars Global Surveyor.
``Our clear short-term goal is to maximize the likelihood of a successful landing of the Mars Polar Lander on December 3,'' said Weiler. ``The lessons from these reviews will be applied across the board in the future.'' [So, is it now NASA policy to specify interface contracts?]