Software Upgrade Crashes UK Air Traffic Control System 233
pitpe writes "Earlier today the computer system controlling most of the UK's airspace failed, after tests in preparation for an upgrade failed. The original failure occurred at the West Drayton centre, which is an old (70's) system, as opposed to the new system at Swanage, which has had its own problems. A system wide reboot to fix the system resulted in the entire system being taken down temporarily."
Software doesn't rust... (Score:5, Insightful)
What WAS the System that crashed? (Score:5, Insightful)
It would help to reduce the coming surge of Microsoft jokes, which is very likely not relevant here.
Re:Software doesn't rust... (Score:1, Insightful)
Because the system now runs, this says the hardware is fine, leaving only the software to be at fault.
Re:Software doesn't rust... (Score:5, Insightful)
What is implied is that its being pushed to its limits. e.g. it was designed for 100 flights a day, when today there are 1200 flights a day.
Those small things which you could get away with before start to become factors in usability and stability.
So what? (Score:5, Insightful)
At least there should be. Computers crash, break, have bugs, etc. They're a tool - a more efficient and convenient tool to be sure.
But when they break, there are contingencies so that planes can still take off and land, and wont just fall out of the sky.
This is also why Y2K was such a bunch of stupidity. We really aren't as reliant on computers as people think. We know they crash and are prepared to handle it when they do.
Re:This wouldn't have happened... (Score:4, Insightful)
This might have happened even if they were running linux. If the software that is used for the air traffic controlling was written badly it still could have crashed.
Re:Links for reference (Score:0, Insightful)
Re:Software doesn't rust... (Score:4, Insightful)
(Which I find strange, cause testing in a system as critical as this should be done in a separate environmnent.)
I assume you've had no previous experience in maintaining a 'vintage' system like that? The code is probably written by a lot of different programmers, each with his own style, poorly documented, and thus very hard to read and understand.
Software doesn't rust, but it clutters up and gets dirty over the years. It won't come apart by itself, but by the hands of a developer writing a necessary upgrade.
Re:More problems... (Score:5, Insightful)
<rant>
I blame Margaret fucking Thatcher, who let the hospitals fall apart and flogged off the viable bits of the infrastructure to her friends (at well below market value). [We're still feeling the effects of this on the railways, which the private sector has run into the ground] Corrupt old bitch.
Re:More problems... (Score:3, Insightful)
Funny, I noticed this about the U.S. system. But I figured it out. It has to do with the fact that civil maintenance is done by civil-service people with a union and a contract, while new equipment and construction contracts with companies are awarded to either the lowest bidder or some politically well connected company. So, the disposable roads/bridges and possibly airports system is a result of trying to contain socialist laborism. And I don't know how I feel about that.
Re:More problems... (Score:1, Insightful)
Golden rules.. (Score:5, Insightful)
Rus
Re:Hang on a second... (Score:4, Insightful)
That's not the way I understand it. From their report, I understand the events went something like this:
- Overnight while it is quiet, new FDP software is brought online for testing.
- Testing was successful, and they brought the old FDP system back online (probably before 5:30 in time for the first arrivals at Heathrow).
- At 6:03 they noticed "errors in the distribution of flight data between Centres." I don't know what exactly this means, but if I had to guess, I'd say that the other Centre was not taken offline during the tests, and some test data leaked onto their live system as a result.
So the problem is most likely not due to a bug in either the new or current software, more likely a bug in the testing procedure they were using.Re:Software doesn't rust... (Score:3, Insightful)
Companies who develop software wanted the release yesterday. Development tools focus on allowing fast development of applications. M$ is taking a lot of criticism on itself by delaying Longhorn.
You can no longer work on a product until 'it's done'. It has to ship, wether it's stable or not. If not, you issue a patch a week later. This is especially visible in games development.
Companies tend to turn a blind eye to some aspects of life when it comes to making a buck, especially if no one's going to hold them responsible for it.
I remember a quote from that horrible movie "Armageddon", where Steve Buscemi, sitting in a space shuttle of sorts, comments that he is sitting on 200.000 parts made by the lowest bidder, $BIGNUM gallons of fuel and a nuclear warhead. Makes you feel save heh?
Who cares if you miss out on your holiday flight? Not the software manufacturer who just cashed in on his ATC-system...
ATC software is scary (aka, Know Your Userbase) (Score:4, Insightful)
The first version of the software was built using standard current interface guidelines and widgets and the testing group that had no experience with older ATC systems were wowed at how simple and yet powerful it was. Pretty much any random person off the street could look at the screen and easily figure out what was going on and how to do various basic tasks. When that version was demoed to the ATC union the union freaked out at how different it was and thus began a cycle of making it more and more backwards.
So, nowadays the next gen ATC software almost exactly replicates the UI of the old non-computerized and semi-computerized systems. On-screen toggle switches and dials, that sort of thing. The FAA and the ATC union have decided that retraining all of their ATCs to use modern computer interfaces would be a Bad Thing. When the computer screen doesn't exactly replicate the interface of the 50+-year-old systems, they freak out and scream bloody murder. On the flip side, kids coming into the field today that have been using computers most of their lives are finding the interface to be counterintuitive to the point of being almost unusable. Middle-aged workers who are both highly proficient ATCs and home computer users report that switching between the two types of interfaces each night when they go home requires conscious effort on their part, since they are so orthogonal.
So who wins? Historical inertia, of course. Why fix the problem today when you can wait for your successors to fix it in 25 years?