Not that many planes crash really if you consider how many are flying all around the world at all times. However when one does crash, we tend to hear about it in the news – it wouldn’t make exciting news to keep reporting “another plane has landed safely” for every other plane! Those that do crash can for several reasons – it could be that the plane wasn’t properly looked after and so something had gone wrong, it could be because a piece of equipment stops working or sometimes it can be because the pilot makes a mistake.
This is Rhys’s area, but I would raise one additional point, which is this.
As we increase the amount of modern technology in a plane (which, these days, is often computer hardware and software), it becomes increasingly difficult to TEST.
The more large and complicated a computer program, the more difficult it is to test and “prove” that it doesn’t contain bugs. This is why Windows crashes sometimes, despite Microsoft’s best efforts!
For a large program, testing under all possible conditions (i.e., all combinations of all values and settings) could take a very very long time, so totally exhaustive testing is not usually done. This means there is always a VERY TINY possibility that the aircraft might behave unexpectedly in a very unusual set of conditions. This happened to an early version of the A320 airbus. However, because there are lots of redundant systems that kick in when one system fails, even if a hardware/software error occurs, it would not normally be enough to endanger an aircraft!
Ian makes good points about software and hardware testing, and it’s true to say that there are immense amounts of software and hardware out there that have bugs of various kinds. This is an interesting example of where science at universities is ahead of the science that’s mostly done in industry. Making complex computer systems error-free is extremely difficult, but computer scientists are making significant progress in that area — however, some companies that need or want complex software systems are not waiting until we’ve solved the problems, they’re just building them and testing them as well as they can,and hoping for the best.
BUT, I don’t think that’s the issue with most air disasters — it’s more that we don’t understand meteorology and atmospherics well enough, as well as how humans and machines interact under stress.
Comments
Ian commented on :
This is Rhys’s area, but I would raise one additional point, which is this.
As we increase the amount of modern technology in a plane (which, these days, is often computer hardware and software), it becomes increasingly difficult to TEST.
The more large and complicated a computer program, the more difficult it is to test and “prove” that it doesn’t contain bugs. This is why Windows crashes sometimes, despite Microsoft’s best efforts!
For a large program, testing under all possible conditions (i.e., all combinations of all values and settings) could take a very very long time, so totally exhaustive testing is not usually done. This means there is always a VERY TINY possibility that the aircraft might behave unexpectedly in a very unusual set of conditions. This happened to an early version of the A320 airbus. However, because there are lots of redundant systems that kick in when one system fails, even if a hardware/software error occurs, it would not normally be enough to endanger an aircraft!
David commented on :
Ian makes good points about software and hardware testing, and it’s true to say that there are immense amounts of software and hardware out there that have bugs of various kinds. This is an interesting example of where science at universities is ahead of the science that’s mostly done in industry. Making complex computer systems error-free is extremely difficult, but computer scientists are making significant progress in that area — however, some companies that need or want complex software systems are not waiting until we’ve solved the problems, they’re just building them and testing them as well as they can,and hoping for the best.
BUT, I don’t think that’s the issue with most air disasters — it’s more that we don’t understand meteorology and atmospherics well enough, as well as how humans and machines interact under stress.
thurkettlek commented on :
thankyou for answering my question 🙂