Hacker News new | past | comments | ask | show | jobs | submit login
The computer crash (timesonline.co.uk)
22 points by tokenadult on June 6, 2009 | hide | past | web | favorite | 7 comments

And next time we have a crash due to obvious human error, the question will be why we spend so much money on technology if it can't be used to save lives.

I'm not denying the observations made in the article - they could be well-founded, and such questions are worth exploring. At the same time, I can't help reading it partly as filler material, and partly as an excuse to bash Airbus (a fashionable choice in some parts of the media, because they're a state-funded enterprise and even worse, they're French).

I'm not carrying water for Airbus, they have had their share of problems. But it's rather unfair to point out that they also have a very good overall safety record. Of course, statistics are abstract whereas tragedies are very very personal.

When people talk about airplane safety issues I like to point out that awesome avionics software, a ridiculous thrust to weight ratio and a great pilot landed an F-15 that had lost it's wing.


The simple fact is bad things happen all the time to airplanes and mostly they just keep on flying. People want 100% safety, but 99.99% is about as far as we can push things when we want to cheaply fly thought the air at 600+ MPH tens of thousands of times a day for hours at a time. Consider the idea that out of all domestic flights on 9/11 less than 1/1000 of them had any real problems.

PS: On a normal day there are 30,000 domestic flights, but most of them where canceled on 9/11.

Agreed. I mean faulty speed readings are obviously a critical problem.

But in infinity 9 cases it doesnt matter - there is an SOP to get round the issue and the plane makes it to the destination.

Similairly multiple electronics failure requiring manual control of the plane probably isnt definitely fatal to a flight (otherwise there would be much stricter faiilsafes, right? look at the eurofighter which has 7 failsafes for some of the core systems - because it phyiscally CANNOT be flown manually - astounding technology btw, but OT :D).

And planes fly through bad storms and are hit by lightening almost daily with no risk.

The combination of loss of auto-pilot in a storm at an altitude that required precision flying - followed by the failure of one of the most crucial instruments involved in that failsafe process is obvioulsy the fault of statistics rather than negligence. In all the airbus planes that have flown for it to happen once is probably highly unlikely - but that doesnt mean it never happens. I mean the chance of winning the lottery is next to nothing - and yet people do :)

(that said I think we can give the article a bit more credence than the one the other day stating it could have been a meteorite strike or something :D)

A less sensationalist (and apparently more informed) analysis can be found here: http://www.weathergraphics.com/tim/af447/

This article seems to be mostly Airbus-bashing. This part about the 'coffin-corner' is just ridiculous. If flying was that dangerous, the safety records should look differently. Airbus' fly-by-wire system is built specifically to keep the plane on the save side of the flying envelope and will actively prevent the pilot from swaying out of it. This is its purpose and design.

More than one accident (the 320 that ran off the airport here in São Paulo a couple years back comes to mind) happened because the computer thought the pilot needs it to do one thing while the pilot was trying to do another. Many pilots regard the Airbus avionics as overly complicated and prone to unexpected behavior since that 320 crash right early on.

During a crisis, the last thing I need is a tool acting unexpectedly.

If sensors are providing conflicting speed measurements, is there a reason not to believe the fastest reading?

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact