Airline software super-bug: Flight loads miscalculated because women using 'Miss' were treated as children A programming error in the software used by UK airline TUI to check-in passengers led to miscalculated flight loads on three flights last July, a potentially serious safety issue. The error occurred, according to a report [PDF] released on Thursday by the UK Air Accidents Investigation Branch (AAIB), because the check-in …
".. but if they were all pro sumo wrestlers then the plane might not have ever left the ground!"
Almost happened to a friend of mine, years ago. She picked up a full load of passengers somewhere in Oceania and then had great difficulty taking off. Later when she walked through the cabin she realised they were the local rugby team - which explained everything. I also noticed in one of my aviation books that the average passenger weight for folk in this region was rather higher than elsewhere.
Another friend was a bush pilot in Australia. He landed at a mining site somewhere in the outback, told the local crew they could load the plane up with X pounds of ore samples and went to have lunch. He too had great difficulty taking off. Eventually he realised they'd loaded his plane up with X kilos of samples. After that he made quite sure he communicated the payload weight more clearly.
Have a safe flight.
I am a little mystified as to why modern commercial aircraft don't have wheel sensors providing at least some indication of total weight including baggage and fuel. And don't they have an accelerometer so they could just set the throttles to "Auto"?
Weight distribution of course is very important.
This post has been deleted by its author
I was going to ask the same thing then I remembered about the fuel quantity indication system on the Gimli Glider. If that is the quality level of aircraft instrumentation then always using a measuring dipstick when the FQIS appears to be working makes sense. If they fitted a similar quality device for an aircraft gross weight indicator then the plane would legally need to be tied down after loading to stop it from floating into the sky like a balloon.
"I am a little mystified as to why modern commercial aircraft don't have wheel sensors providing at least some indication of total weight including baggage and fuel."
Those sensors would only give a reasonable indication if there is absolutely no wind. The minute you get any wind the wings will develop lift which will reduce the load on the wheels (while the mass of the plane remains constant), leading to an erroneous reading.
Happened with a friend of mine around 20 years ago, she was in her seat already, the back row next to the lavatories, one of the last to board, and here goes the announcement:
- Miss V... Rei.....t, Miss V... Rei.....t, please proceed to the exit, our flight is overweight.
By coincidence, my friend was not a small lady and the walk of shame along the whole length of the isle was something she said she would never ever forget.
ok, how about weight sensors in the seats then? Even an approximate value would justify adding a few coins to the cost of the seat... [I've been looking at price of pressure sensors lately, and they're not that expensive and often use I2C]. Extra added bonus, you'd know if the seat is empty at any time during the flight.
Messing with how seats get mounted is a big deal, the seats and seat attachments are a key safety feature. Any changes will be costly.
All those sensors need to be connected to something to read the measured value. That's a lot of wiring.
The sensors, accompanying wiring, measurement units, and communication to flight controls are going to need to be tested and certified (they're flight critical, wrong weight = you could hand a very bad but very short flight).
I'd estimate $500/seat, installed price. That's after amortizing in design and certification testing. I have no experience with FAA part 121, so I could be well off the mark.
This post has been deleted by its author
Some of them do, but it tends to be for verifying the centre of gravity instead of the gross weight. I think you'd struggle to measure the weight with sufficient accuracy to prevent an issue like this. An Airbus may have spotted the problem but only when airborne.
Take-off performance monitoring is something that major manufacturers have been struggling with because although it seems superficially easy, it turns out to be really complicated to assess other than in the gross sense, which wouldn't have been sensed in this situation.
IMHO since the problem was known about at the airline a second check should have taken place at the gate when boarding cards were scanned (assuming a human is still involved).
Like a weight scale embedded in/under each seat to determine exactly how much passenger weight is aboard, scales in the cargo hold to determine exactly how much cargo weight, etc. This would also give the pilot advanced warning if either source suddenly shifts & risks unbalancing the plane. It's not like it's rocket surgery...
measure passenger mass automatically.
The obvious answer is not to try to fit something to the aircraft but to fit something at the boarding gate where people have to walk through. A simple weighing plate next to the check-in desk would work for individual passengers, though a bit more awkward for a group booking in together I suppose.
Wouldn't be 100%, but certainly more accurate than assuming one weight for children and another for adults.
Seats could even be allocated on-the-fly to deal with centre of gravity issues.
M.
Could they add weight sensors to the body scanners used at security?
Give the passengers an RFID tag that identifies them to the scanner, add the weight, feed the data to the airline to match the ID to the intended flight and total them all up to get total passenger weight.
Give the passengers an RFID tag that identifies them to the scanner, add the weight, feed the data to the airline to match the ID to the intended flight and total them all up to get total passenger weight.
Good idea. Then the passengers could be electronically stunned - and moved towards their flight on conveyer belts. And loaded in by the baggage handlers, in a much more convenient manner.
With respect, yes we do.
Barring the 737-Max, which was applying changes to an ancient grandfathered design, the design of new airplanes including their automation is one of the most rigorous design process out of any industry. Only the nuclear industry is comparable to the same safety standards. The design and implementation of automation in new designs, including software, goes through an insane amount of detail and checks. When people ask why airplanes manufacturers don't implement some piece of fancy technology (which must be easy right?), the answer is normally that they need to prove it safe first which is a much longer process than it would be in any other field.
And rightly so. I can't think of any other industry with such a good safety record, and it's a hard earned safety record. The industry as a whole has been at the forefront of safety for the past 50 years, and is something other industries and sectors look to learn from. The failures of the 737-max just show that complacency and resting on your laurels is deadly.
Reminds me of the quote from TUI stating the safety of their passengers was paramount, after also advising that the problem wasn't immediately sorted out because of the holiday break.
Seems that statements claiming a gold standard approach are abound after every incident, demonstrating the tongue of the PR department, is truly bifurcated.
If systems are thoroughly checked and double checked, by highly skilled engineers, then I need to shove my CV under some noses and get, what must be, the cushiest job on the planet....signing off inspections from the comfort of the pub.
I'm sorry, but re-read what happened. That's the sort of error that should have either grounded the fleet, required some sort of fallback to a manual process, or overridden the computations to assume all passengers are adults.
This was a callous, even cynical decision by someone in the airline to go for profits over a blatant safety issue. Full stop. At Boeing, it can at least be argued that no one person had all the data to know that they would be endangering lives.
"Weight distribution of course is very important."
I actually wrote check-in and seat allocation software (back in the 1980s, I think) which was mostly used by small, regional airlines in rural Africa. It was necessary to distribute passengers evenly around the cabin to keep the small planes balanced.
Fundamentally this is because the sensors would measure weight, but what you really need to know is the mass of the plane. The sensors readings will vary according to wind (which will provide lift while blowing over the wings even went the plane is stationary), the load distribution, and many other factors.
The issue is also not only about the mass, but the mass distribution. Having all the mass at the front of the plane will present a lower angle of attack, reducing the lift generated for the same amount of throttle input and making it harder to take off. Having the same mass further back will present a higher angle of attack, making stalling more likely. This is why we like you to sit in your allocated seats.
The throttles are set to auto - they use the mass of the plane to calculate the take-off thrust, but the mass has to be entered by the pilot.
Hmmm... Requiring Boeing to do some calculations based on sensors fitted to a plane... I seem to remember they've had some issues with that sort of thing in the not-too distant past? Maybe best if they don't try anything clever just yet. At least, not until there's a grown-up available to check their work!
I flew from Kathmandu airport to Lukla 'airstrip' in Nepal (before the extension to the Lukla strip). The Twin Otter aircraft had a stick that the pilot hung down from a hook at the rear of the aircraft. He would only attempt take off if the stick did not touch the ground.
I'm not sure if landing at Lukla (and hoping the aircraft stopped before hitting the solid stone wall a the end of the runway) was more or less exciting than take-off (and hoping the aircraft achieved actual flying speed after literally dropping off the end of the runway before hitting the bottom of the valley below). Possibly aiming for the 'notch' in the mountain ridge half way between Kathmandu and Lukla was even better.
"I'm curious why a bush pilot in Aus would use pounds?"
As the previous poster said, aviation - especially if flying an American designed light aircraft, uses a whole load of different units. Pounds for weight, inches for centre of gravity datums, potentially US gallons for fuel quantities. And the units used are specified in the aircraft manual which is a legal document so everything needs to be converted back to those units to make sure you're not over the max weight and the CG is within limits.
So you have to convert your USG fuel required into UK G or litres - and the potential for weight-affecting errors just here is massive if you're not on the ball as it's not volume you need, it's the actual weight...
The units in use simultaneously on a modern aircraft:
- feet
- metres
- knots
- kg
- metric tonnes
- mach
- litres
- quarts
- hectopascals
- inches of mercury
- pounds per square inch
- nautical miles
- degrees celsius
- degrees magnetic
- degrees true
- degrees grid
- TPR (turbofan power ratio)
- kilowatts
- megahertz
- kilohertz
> told the local crew they could load the plane up with X pounds of ore samples
Doesn't Australia use the metric system these days? At least for measuring anything where accuracy matters. (Wikipedia page https://en.wikipedia.org/wiki/Metrication_in_Australia says metric became the only official system in 1988).
It is possible the local crew had balances only marked in Kg.
I used to fly regularly with one of two adult female colleagues (each was "Ms"). One barely came to 50kg, the other wasn't much below 100kg. The whole idea of assigning a standard weight to each of the 200+ passengers on a plane and assuming it's sufficient to calculate any aspect of aircraft performance just seems daft.
It's not pretty pointless, It's very pointed.
Just because you are not affected this time doesn't mean you couldn't be in the future.
Stop thinking about just yourself and try and grasp the possible effects of this error.
It seems people were doing their jobs diligently so well done to them, and the way they had been trained.
That's how aviation is supposed to work.
There are many vital links in the chain, the pilots are the most visible in any enquiry of course.
But pilots can only deal with what is in front of them and they rely on the supply chain being accurate.
TUI failing to respond to the specific question El Reg asked about where the system programming is interesting, I think I'll ask them that question myself and see if I get a response.
@Foxglove: "TUI failing to respond to the specific question El Reg asked about where the system programming is interesting, I think I'll ask them that question myself and see if I get a response"
Don't hold your breath. TUI is a German-owned tour company which bought out Thomson Holidays some years ago and continues to uphold the previously long established Thomson tradition of poor service, arrogant management and indifference to customer complaints.
That it should in the face of this situation glibly issue a Press Release saying how much it values the safety of its customers and staff says all there is to say about corporate mendacity, as well as TUI's belief that it can get away forever and a day insulting folks' intelligence with nonsensical corporate PR bilge-speak.
(Which it probably can, given that the overwhelming majority of its customers are unlikely to even know about this load issue, or understand it, or ever get together to boycott the business.)
"TUI failing to respond to the specific question El Reg asked".
Quite frankly I don't think TUI should bother to answer that question at all.
I had a small software company for ten years and I can assure you I would never have pointed to one programmer or an other to a customer with a question like that.
I think some people should start asking themselves the question - "why do I ask that question". What am I going to do with that information, and why do I think it's important to me.
I think some people should start asking themselves the question - "why do I ask that question"
To probe whether they are aware of the problem that question is highlighting, and more specifically, whether they would be answering that question now or several closely related ones in front of a board of inquiry.
Plane accidents tend to invite those questions.
As noted, TUI is a German company, so I have a hunch that the programming may have been done by a programming team in Germany (as software engineering is still treated as a reasonably respected profession there (whereas a British company would just have outsourced it to one of the usual suspects, who in turn would have outsourced it to whichever developing country wet-behind-the-ears newbie graduates they could get cheap that month)).
Possible further evidence for this hypothesis is that although "Fraulein" technically does mean "Miss" in German, these days it only tends to be used as a form of address for girls, rather than adult women, and so you can see where a cultural misunderstanding could arise.
(But this is only speculation on my part.)
"TUI failing to respond to the specific question El Reg asked about where the system programming is interesting"
Not really. The Germans could hear the dog-whistle and weren't playing. The Reg does have a few writers who seem conspiratorial alt-right loons, so it's pretty clear what the 'question' was for.
Obviously the programmers were actually German, not those icky brown skinned types the writer was attempting to blame.
Obviously the programmers were actually German, not those icky brown skinned types the writer was attempting to blame.
Highly unlikely the programmers were German as those are thorough enough not to take a title as indication of age, especially not in the presence of other indicators as ticket class (child tickets) and date of birth. Besides that, Germans have a tendency to err on the side of caution.
Down vote for "icky brown skinned types".
I spent many decades doing corporate software development, and there is ONE place in the world where projects like this are programmed, if not in the home country. And it is not Germany. The problem is not "cultural differences" but incompetence in management styles at both ends of the arrangement.
Standard response from a neo, accuse others of racism.
What's a neo? Neo-liberal, neo-nazi, neo-from-the-Matrix, neodymium?
Also, come to think of it, what the fuck does dogwhistle mean? Is it one of those things like gaslighting that you can just sprinkle into your sentences to imply that somebody you disagree with is bad - because you have no actual fucking evidence to back that up if you were to claim it in an actually comprehensible way?
Perhaps we should say what we mean, and assume that others are doing the same. Rather than deliberately obscuring our meanings, while at the same time claiming others are saying things they haven't actually said.
Rant over. Thank you for your patience.
I would think it would help safety to know this as other companies may also be outsourcing safety critical software to the same country. Knowing the assumptions that a particular set of programmers may make would likely be a good thing.
If past aircraft accidents have taught us anything, it's that they are caused by a series of small events or defects all coming together. The holes in the Swiss cheese lining up as it were. Knowing that developers from X country will make this assumption is the kind of thing that would be taken note of in the industry. Doesn't mean they won't outsource, it just means they will be more precise with their specs.
It just closes one more of those tiny holes. So I'm not sure keeping shtum about where the software was developed is really in the spirit of openness that air accident investigators try to foster.
I think you should probably read my comment properly, it don't think it says what you think it does.
You are correct that I don't care about gender or religion, which is why I specifically didn't mention them. The developers were likely a team that consisted of a variety of genders and religions (hopefully). The issue at hand here is cultural and likely shared by everyone in the team.
I don't talk about banning anyone and even say that it doesn't mean companies won't (and I meant to imply shouldn't) outsource to this country. They still should if they want to, just make sure they specify more precisely what they need in order to avoid these cultural differences causing issues.
I only say that the industry would likely be interested to know the country where the code was developed so they can ensure that any specifications for development sent to that country in the future avoid this error again. I don't care where it was developed, it could be the US, UK, India, China or Timbuktu for all I care. The fact is that the spec for this development didn't take into account the cultural differences between where it was specified and where it was developed and therefore this error crept in. Disseminating this information will prevent it happening again.
The total weight of the aircraft is not the only factor here; the distribution of weight enough to alter the Centre of Gravity to a point where the trim cannot bring it back to flyable limits could also come into play.
Imagine a situation where of those 27 passengers, 15 were a Ladies choir and all "Misses" - so were thought to be 35kg each but were in fact 69kg - and they all ask to sit together down the back, or at least behind the CofG enough to pose a danger but not enough to tip the aircraft on its tail during loading (that happens!!).
The Weight and Balance system would allow this at check-in based on this erroneous coding, pilots would be none the wiser too. At rotation, rotation can't be stopped at 10 degrees, continues to the stall, aircraft then pancakes.
Lots of bad things lining up, but the Swiss Cheese failure mode is frequent in aviation.
No... it would likely compensate by seating the males (average weight for W&B calx = 83kg) and anyone who didn't care where they sat, forward of the CoG.
And a small correction to my original post:
"alter the Centre of Gravity to a point where the trim cannot bring it back to flyable limits"
..should really be:
"alter the Centre of Gravity to a point where the trim cannot bring the Centre of Lift (or Pressure) back to flyable limits"
They were rightly worrying as that one being has lately developed a tendency not to pick up that hot line (or any other line for that matter). As I understand it, the chances are slightly better when you call the secretariat, she was last reliably known to pick it up about a century ago for some children near Fatima, Portugal.
Why use title rather than date of birth - pretty sure those details get passed on or you at least are paying for an adult or child ticket in which case age is normally required. Less of a misunderstanding, more a case of simply running out of talent.
Not that either is a good indication of weight mind you.
Why use title rather than date of birth
The Title field was probably at the top of the table and the developer didn't look any further.
When I took over my current role about a year after an ERP implementation there was a VAT report that never balanced. It had been referred back to the developers repeatedly with no solution found. I took one look at it and spotted the value*0.15 calculated field for the VAT, despite the fact the actual VAT charged amount was in the table being reported from.
Some developers are idiots.
@dak
Perhaps if the software for all TUI airlines including TUI Airways was produced in Germany and perhaps also tested on the German TUI airlines like TUI fly Deutschland first, even well.
But then later used in Britain, this misfortunate misunderstanding with the "miss" was revealed.
Not a good system design and indeed one would hope there was more reliable methods to get the weight of passengers.
Indeed, even without the specific mistake made here, the whole thing seems like a really bizarre setup. You already have the information you need, so why not just use it instead of trying to reproduce it from a different piece of information that may not even be well correlated.
On a related note, it's always struck me as rather odd that airlines are always so eager to weigh checked luggage down to the gram and charge extortionate fees because weight is so important to the ability to fly, but they don't care in the slightest how heavy the actual passengers are, or how heavy the rest of their luggage is. Weighing passengers has obvious PR problems because fat people will complain that the laws of physics are discriminating against them. But even without opening that can of worms, I could check a bag full of helium balloons into the hold and take a bag full of gold bars as my cabin bag and no-one would bat an eye, but if I swap those bags around suddenly it's a huge problem because of the excessive weight.
It just seems odd that something so important to both costs and not falling out of the sky is left essentially to pure guesswork. On average, you'll usually be close to correct. But it only takes one plane full of non-average people to result in hundreds of deaths. You'd really think it would all be taken a bit more seriously.
Having seen how much stuff some passengers want to shove in the hold, I suspect the limits and excess charges imposed here are as much about ensuring people don't take the piss as they are about ensuring a safe flight. After all, if you can afford to pay the excess charges, you can still take all those extra suitcases, packing crates, shrinkwrapped household appliances etc. with you, so clearly their weight isn't a concern unless you're really pushing it or have found yourself sharing a flight with another passenger who's also trying to move house on the same flight.
In contrast, hand luggage is somewhat more self-policing given that you have to lug it onto the flight yourself. You're always going to have the odd edge cases (such as your bag full of balloons and a carry-on full of gold bars example), but in the main it's a hell of a lot harder to carry excess weight *into* the cabin, than it is to load it onto the conveyor at check-in and have it packed *under* the cabin.
I used to fly regularly to an airport that was high, hot and short. I can assure that hand luggage was weighed, and quite often they would only load a portion of checked baggage - which would either fly in on a later flight, if you were lucky, or more often get flown to the low, hot and long runway 6 hours drive away, and get shipped up by lorry.
All airlines have hand luggage weight limits - they tend to be enforced ,as ChrisC says, by making you carry the stuff - but I've had mine weighed quite often in recent years, more often on short haul flights where it may well be that the plane has less margin.
Exactly, passport gives age and gender and I am anticipating that with the increasing costs and reducing margins of air travel it is only a matter of time before scales appear in the checking in process, which would catch the Ryanair tactic of wearing a big overcoat with fully loaded pockets :-)
I used to know a German lady who lived in the UK but wintered in Spain. She could have been the origin of the stereotype of a loud overbearing German. On one of her trips to Spain, she did the usual trick of turning up at Heathrow with five suitcases/bags and it was suggested that she needed to pay an excess baggage charge. She looked around the check-in hall and pointed out some fat guy checking in with one small bag and said "Is he paying an excess charge? You can weigh him and his luggage and me and my luggage - if I weigh more, I will pay. If not, I will not pay". The lady at the check-in desk decided to forego the excess baggage charge.
There is something to be said for weighing passengers and all their luggage. I've often seen people checking in and declaring a single carry on bag, while wearing a rucksack under a coat or having a friend or family member look after addtional undeclared carry-on bags.
JohnG,
I used to know a missionary who was returning from New Guinea in the 90s. I've had dysentery and lost 50 pounds. As I paid for this return ticket when I weighed more, you should let me bring two suitcases back free to make up for it.
I still know several missionaries, and they are all experienced and talented excess baggage blaggers. Including one who got a brand new washing machine from Heathrow to Lagos as well as an extra two suitcases without paying. But flights to Africa always go with loads of unpaid excess baggage - presumably because the airlines can't sell the hold space for air-freight - and know it's easier to put a bit extra on the ticket prices than argue with every single passenger.
Why use title rather than anything else
I think the obvious elephant in the room has been avoided. And, here comes Captain Obvious [me].
If an entire American football team decided to self-identify as 'Miss' on their next flight, then this whole algorithm would be thrown on its head. Sorta like a team of elephants self-identifying as "mice".
(how come nobody else said anything like this?)
[troll icon, for obvious reasons]
weight sensors under the seats makes a lot more sense. Who knows, they might pay for themselves over time with better takeoff performance and fuel cost reduction.
Once upon a time the pilot of an aircraft would have to be responsible for a pre-flight check including air-worthiness of the plane, knowing takeoff weight, fuel amount, and maintenance history. They would also have to submit a flight plan. All before takeoff. It was part of being a pilot, and not necessarily anything to do with knowing how to actually fly the plane in the air. Several disasters later we know that its simply not possible anymore for the pilot to gather all the info they need to know. So its then the responsibility of the airline to ensure that the required checks have completed and information is accurate. Pilot error is often attributed when things go wrong, but it should be obvious from the above that modern pilots of larger aircraft are completely blameless when things like takeoff weight are wrongly calculated.
a women's weight, a woman's height, a man's height...
No opinion, because I'd rather point towards the history of collecting data to create averages for chunks of the population. Fascinating slices of life.
From the USA gov needing to know what size of military uniform to produce informing our current S M L XL labels, to percentile of pop height range a car driving seat must accommodate... and doubtless many edyfying case studies I've not read.
It is odd that many folk of average height (for their group, obviously) in the UK ( c. 6") are sold duvets that ask them to choose between warm toes and chilly shoulders, or vice versa.
When I was a teenager, 40 years ago, the biggest shoes you can buy in the shops were 12s. My feet eventually grew to 10. My sons' feet are 13 and 14. The biggest shoes you can buy in nearly all shoe shops are still 12.
I also note that a modern car still fits in a modern car parking space, but now you have to exit through the sunroof. Which you almost certainly don't have because you will know certainly have air conditioning!
«It is odd that many folk of average height (for their group, obviously) in the UK ( c. 6")»
Umm, 6" is a rather different sort of average 'height' than the one you're referring to! If you want to use squiggle units, it's 6' that you mean! (I'll happily stick with cm, thanks!)
don't ask me what happens to someone whose 16th birthday is while in the air
Age at boarding (and take off) on departure is used (and there are very few flights that board before midnight and take off after, so in case there is a discrepancy, the passenger gets the benefit and congratulations).
From the article: "On July 17, the developer(s) working on the check-in application "adapted a piece of software, which changed the title of any adult female from Miss to Ms automatically."
So if they don't have the passenger's date of birth, how are they able to change the title of adult females in the application. And if they do have date of birth, why didn't they use this for the weight calculation?
The only conclusion that I can draw from this is that they have some exceptionally dumb software developers. But then, pretty much every airline website I've ever used appears to have been created by exceptionally dumb software developers, so perhaps this is standard for the airline industry.
Not acceptable.
You do not program for how your own country interprets values. You program for how your client country interprets values.
Mistakes like that lead to losing a satellite because you transmitted data in metric values while the program treated them in imperial values.
TUI are a holiday airline so AFAIK, every flight is international and requires full passport info to be provided, including DoB.
On the other hand, taking an average weight based on purely on a single cut-off point probably isn't all that accurate anyway. Some teens are "child" sized while other of the same age are easily "adult" sized.
I remember my first ever flight on a school exchange visit to France. The first leg of the journey was on a relatively small 3 engined jobby (Trident?) to Heathrow which was 2+2 (2+3 maybe) seating. single aisle. With 40+ kids on board ranging in ages from 12 to 16, the cabin crew took care to rearrange where we sat based on size of each kid.
Thanks, John. I have flown out of the UK internationally a number of times, but have never booked the ticket from within the UK, so apologies if this is a standard UK arrangement. I really don't recall giving this info when booking elsewhere for international travel, but it's been a while.
I'd imagine the airline approach to determining passenger weight, if based on age, would need to be standard if they also flew domestically and passports didn't figure; that's not to say TUI couldn't rely on this if all the pieces of the puzzle are there. If they are, why even look at passenger titles?
This post has been deleted by its author
I was involved with the Epson EHT-40 touch screen handheld computer that was used by the Concorde baggage loaders. It calculated how to spread the baggage based on the seating plan and customer weight (scales at the checking?).
The mean weight of passengers on the trans-Atlantic route, was higher than any others.
In the course of browsing this thread I've checked EasyJet, Aer Lingus, Ryanair (I know), BA (also, I know...). All have an adult ticket category 16+ and infants (<2yrs), all except EasyJet further divide <16 into ~12-16 (teens/young adults) and 2-11 (children).
""This should not have stopped the program from working, but as this was a 'fix,' it could not be known for sure."
This is what screamed at me... you don't test your fixes Tui? Really? What do you do, CI/CD hack them up to production and see how they perform there? If the software doesn't crash, fingers crossed the plane won't either?
Testing? We don't fly there...
...that normally on takeoff, the pilot would be using everything short of War Emergency Power, throttles to the stops. I would hate to clip an obstacle at the end of the runway because the airline was trying to save a few bob's worth of fuel, and the passenger cohort had more than the normal fraction of bloaters.
Every day's a school day.
Noise pollution - that brings back memories.
Standing in the carpark at Heathrow waiting for a bus into the CTA when Concorde took off - or an MD-111.
Always left the car alarms blaring away as soon as the engine noise died down. And they weren't even hot-rodding!
Stuck in a traffic jam on the M25, at the end of the Heathrow runway, windows wound down because it was blistering hot, when I thought the world was coming to an end...
Concorde goes thundering overhead at a few hundred feet, giving it the beans.
Scared the living crap out of me. The word LOUD doesn't really do it justice.
I remember also being at Reading rock festival in the early 80s, and it used to drown out the music every afternoon as it climbed out from LHR.
Wish I could have afforded a flight. What a plane!
I would have thought......that normally on takeoff, the pilot would be using everything short of War Emergency Power, throttles to the stops. I would hate to clip an obstacle at the end of the runway because the airline was trying to save a few bob's worth of fuel, and the passenger cohort had more than the normal fraction of bloaters.
No.
Yer average commercial passenger jet's engines can suffer catastrophic consequences if they spin too fast (overspeed), or if the exhaust gas temperature (EGT) gets too high, or they try to push too great a mass of air. You need sufficient thrust to reach Vrotate before you run out of runway (and preferably V2 soon after), but importantly, not so much that you exceed the speed rating of your tyres before lifting off. This means you need to carefully control your thrust to ensure you accelerate enough to reach Vrotate before the end of the runway, but not so much as to exceed the tyre speed rating. The margins can be surprisingly small.
More background details here:
Turbofan Technology:Jet Engine Thrust Ratings
Boeing Aeromagazine: Exceeding Tire Speed Rating During Takeoff
BAA Training: Did You Know about Aircraft Take-off Speeds: V1, Vr and V2?
Cockpit News: WHat do EPR, EGT, N1 and N2 mean
Skybrary: Engine Pressure Ratio (EPR)
So, for my non-aviator understanding, the limit is that you need a large enough time window between what I'd think of as take-off speed (Vrotate) and the tire speed limit (hopefully lower than Vr for still days) to actually take off? And accelerating too fast would be an issue for that.
I sure hope the tire speed limit is a nice margin above Vr, otherwise take off will be with burst tires, which will result in a not so smooth landing.
In the linked Boeing document above gives a case study showing:
Scheduled Ground Speed at Liftoff: 199 knotsRated Tire Speed: 204 knots (235 miles per hour)
and goes on to say
...case study showed that a rotation rate that is 1 degree per second slower than normal can result in a 4- to 5-knot liftoff speed increase. This is in addition to the increase in all-engine takeoff distance associated with the slow takeoff rotation (see fig. 3). This illustrates how a slower-than-normal rotation rate can easily use up what may seem like a large tire-speed-limit margin, especially if paired with a higher tailwind component than accounted for in the takeoff analysis used for dispatch.
Some operators have elected to simply examine the tires after an overspeed takeoff event using the normal tire inspection criteria in Chapter 32 of the Airplane Maintenance Manual. if no damage is found, the airplanes are dispatched normally and no further maintenance actions are performed. Based on many years of service experience, this approach seems to have worked well because very few, if any, tire tread losses have been attributed to an overspeed event. Based on this service experience, Boeing has typically not objected to this practice even though there is no overspeed takeoff capability specifically designed into the tire.
They use power appropriate to the situation, and monitor how the aircraft behaves. The pilot can use extraordinary measures to keep the plane safe, if necessary. If normal operation uses extraordinary measures, then we should question why.
Here's an example: if you're in danger of clipping a ground obstacle, use "max angle of climb" aka Vx. Basically, throttles to the max and tilt up as far as you can. You're high and slow, and if an engine burps, you die. This is not normal!
Also, this is about more than just power. If the aircraft is too nose-heavy, it won't take off. If it's too tail-heavy, it won't stay airborne. If your university women's rugby team fills the back half of an Embraer E170 and were listed as "child" then maybe they put heavy cargo in the tail to balance it out. The result could be noisy and unpleasant.
Bear in mind the aircraft is designed to be able to climb away from a take-off even if an engine fails at the worst possible moment so in an emergency they can produce a lot of thrust. You don't want to do that most of the time for all the reasons mentioned above, but if it's a choice between that and crashing...
"Bear in mind the aircraft is designed to be able to climb away from a take-off even if an engine fails at the worst possible moment so in an emergency they can produce a lot of thrust."
Not only that, a properly designed airline engine will have multiple redundant sensors for critical things like pressures, temperatures and speeds, also throttle lever angle or equivalent, and they will be reasonably accurate and reasonably trustworthy. Guessing (or "modelling" as it's sometimes known) any of these numbers is often a bad idea, especially when the topic of concern is safety related, e.g. transient overspeed/shaft break protection in a 3-shaft engine where only two shafts have actual speed sensors, especially if the real engine behaviour in extreme circumstances isn't a good match for the model.
Regardless, this TUI incident does seem to shed some light on the software development practices (including specification and testing and the response to anomalies) of the companies involved.
I'm deliberately ignoring the two-faced PR spin.
"It's suggested this won't happen again"
What won't happen again?
This particular problem - maybe not now it's fixed unless, of course, somebody reverts to an older version or looks at the new code and thinks "That's not right for a child"?
Or developing against an insufficiently detailed spec that assumes culturally specific knowledge on behalf of the developer?
The system programming was not carried out in the UK, and in the country where it was
performed the title Miss was used for a child, and Ms for an adult female, hence the error.
A manual solution for correcting the problem was quickly identified that involved a team
identifying upcoming flights, checking each booking, and changing all adult females with
the title ‘Miss’ to ‘Ms’, which overcame the problem. Subsequently, this work was shared
between two teams, and the process was completed every afternoon and evening for the
next day’s flights. It was checked again every morning, where possible, before flights
departed.
But the point is, that anyone with half a brain should have thought - hang on a sec, that's not going to work....
And to get round it with a manual check is just dumb.
As someone else said - why not just use the date of birth of the passengers? Come up with a lookup table of average weight vs age, do the calculation, add 10% to err on the safe side and bob's your uncle.
I mean, honestly....
Or - even better. The bags are weighed as they are checked in. Why not weigh the passengers as well, carrying their hand luggage? How long would it take? Probably no longer than the security check....
Long ago used to fly passengers in hot air balloons,. We had a weight limit of 200 lbs per person. Can't think of a single person every paying the over-weight premium and we didn't even have a scale to check. Wasn't critical because the balloon wasn't going to fall out of the sky, but I used to put a glove over the instruments so I couldn't read the temperature with some on board.
I got two free rides on a helicopter at a fairground back in the 90s. A bunch of rather overweight older people were lined up, all weighed, and no calculation would come up with the value that the pilot specified (certain weight, certain number of people arranged in a certain way).
The flight assistant spots me, calls me over. I'm the right weight. So I went up twice (with each group) to make the weight correct. They paid something like fifty quid each, I paid nothing. Sweet!
I never really understood that. I can see not wanting to be too heavy, but don't get why it was equally bad being too light. After all, their journey home won't be with a bunch of people...?
Possibly a weight distribution problem at that point, rather than a maximum weight problem - if you had one of the larger passengers sat on the left (say), then you'd need enough weight sat on the right to maintain the correct balance, but if none of the other paying passengers were then light enough to provide that balance without exceeding the overall weight limit... Enter your good self, light enough to avoid causing the latter problem, but heavy enough to correct the former one.
This post has been deleted by its author
Could be said about just about any crash 'hit the mountain' vs 'didn't read the map' or 'Wings fell off' vs 'didn't tighten up bolt that held wings on'.
Still you are right - but there are lots of good reasons why you might not take off with full fuel tanks, apart from economics, e.g. overall aircraft weight limits, weight limit on landing, runway length, altitude, air temperature and I'm sure loads of other factors.
How many crashes and deaths have we got so far due to lack of enough fuel on planes?
One pilot managed the world's longest glide in a passenger jet but that was down to a different problem. Although maybe that should be two problems. The main cause was poor maintenance leading to a fuel leak the other was making a decision based on inadequate information.
The last bit of piloting was a nice bit of flying though.
a 120km glide. Scary.
Well luggage is weighed at check in so have a set of scales embedded in the floor by each departure gate big enough to weigh the average family and their walk on luggage when they present their passport and ticket for checking. All it needs to do is keep a total of the weight passing over it which can be transmitted or communicated to the plane.
Incorporate a weighbridge into the airbridge (or flight steps, if that is in use). I have seen people carry hand luggage onto a plane which weighs considerably more than a small child does (judging by the effort needed, often by more than one person, to get it into the overhead locker).
FTAOD: that's the carry-on case they're man-handling into the locker, not the small child...
69kg's ? have you looked at your average couch-potato these days ? they better have a 100% margin of error on these calculation programs. Fancy a flight full of pensioners in their floral-curtain dresses like 'Violet and Daisy Bucket' and their pot-bellied 'Onslow style' husbands ... not a chance that falls in the 69 kg range.
We all need to go through those body scanner yes ? Have a built in scale in those. Walk up, insert your ticket , As you are being scanned you are being weighted. your ticket comes out the other side. The scanner takes a snapshot of your ticket and can correlate the weight to the ticket. the computers do the rest . Anonymous as nobody gets to see the individual numbers except the total for the flight by the pilot.
That can't be right. Even if the plane were sitting out in the middle if a field, rather than at a terminal building, and the wind happened to come from perfectly head-on, it wouldn't generate significant lift.
I think it's more likely that weighing equipment of sufficient precision is surprisingly hard to build, expensive, or heavy. And this isn't a calculation that's hard to do, or needs to be particularly precise, so there's no point.
People seem to be missing the bit where this is our hypercautious air safety system noting a flawed procedure which came close to breaching a limit that is itself very conservative. No planes or passengers were endangered.
So we are looking for a system that is > 99% accurate (the flagged issue was 1000 kg overweight on a 100 ton aircraft). You are probably weighing an integrating at 3 points. You need to factor out wind, because I can ensure you that a head on wind would generate some lift - but I guess you could either fit 360 degree wind speed monitors or do a 360 degree loop on the taxiway. So then you have a number - do you feel lucky? All 3 sensors working, calibrated, accurate? Maybe we need 3 sensors per wheel so we can do a majority vote on the sensors?
Isn't this the fault of the analyst?
Someone had to spec the system and present it to the "client" who would then sign off on the development.
I think the main difference to my day is that we would have tested the software properly before making it live.
They don't seem to be able to do that nowadays.
We did it by intuition, knowing what had to be done to achieve a 100% working solution, rather than scrambling with agility or whatever useless dogma they are currently promoting qualifications in.
You can't replace good programmers with processes and average programmers. The result is not the same.
"We did it by intuition, knowing what had to be done to achieve a 100% working solution,"
I call bullshit. I've been writing software for a living for more than 40 years now, and there has never been a time when there was any software without bugs.
Icon: Me.
A piece of software with NO bugs.....
10 PRINT "Hello World "
20 GOTO 10
Prints Hello World all over the screen continuously, exactly as intended therefore no bugs....
Making definitive statements is hard as there is always a case to disprove the statement.
"A piece of software with NO bugs.....
10 PRINT "Hello World "
20 GOTO 10
Prints Hello World all over the screen continuously, exactly as intended therefore no bugs...."
All over the screen? You missed a semicolon at the end of line 10. There you go, found your [first] bug.
But surely this highlights the crux of the problem - requirements defined that are open to interpretation?
The OP may have a different understanding of what "all over the screen" means to you or to me. His interpretation was to print "Hello World" all over the left hand of the screen. Yours was to print it serially all over the screen. Someone else may have meant "Hello World" to appear randomly all over the screen. One person's desired result is someone else's bug or 'feature'.
If you outsource your programming to someone/somewhere else, then your requirements have to be clearly defined. If its left open to interpretation then things like this will creep in. Same with the Miss/Ms issue.
Also, if the person writing the code is inexperienced in the field they are writing the code for, then that person may not know to question the requirements spec. They will follow their *interpretation* of the spec. I've seen that on more than one occasion.
1980's Basic v 2020's Python (If my aging memory serves me well*).
So easy for an error to creep in even here.
*Been teaching myself a bit of Python - haven't written any code for about 30 years. The semi-colob at the end of the line caught me out so often!
...there has never been a time when there was any software without bugs...
But it used to be the case that the attitude was - yes, there are bugs, bound to be - but let's try to make sure there are as few as possible, and they are as trivial as possible.
The attitude these days seems to be meh - bugs - who cares?
This particular issue is incomprehensibly bad. To my mind, whoever designed this software and whoever let it get into production is as guilty of gross negligence as an electrician who wires up a plug wrong. They should be sacked.
Tui have not only outsourced the programming, they've outsourced everything. They have no idea how things are designed and don't have the expertise to test the result.
I wouldn't leave it to intuition, because if you leave it to intuition you get stuff like we've just seen.
Also, there are tables of average weight by gender and age. With the title and DOB (which you have to give when filling in passport details) you can work this stuff out.
When one of the Canadian airlines was switching from English to metric units the fuel was loaded to the correct value but in pounds not kilos. The plane ran out of fuel in flight and had to glide to a landing on an abandoned RCAF airfield being used as a drag strip. Outside of a few bruises there were no injuries and no fatalities. This type of error is far more common than is realized.
fuel was loaded to the correct value but in pounds not kilos
Not quite. The fuel gauge was inoperative, and the existing fuel load was measured using a dipstick before departure. The depth→volume calculation used the wrong units, so the amount of additional fuel loaded was insufficient for the flight.
There's no relation between this case and the GG. This plane wasn't short on fuel. The throttle wasn't opened far enough, that was all. And not to save fuel, either, but because take-offs aren't done at max throttle.
The GG is practically a unique case. There have been vanishingly few airliner crashes due to fuel exhaustion at all, and all the others I'm aware of were basically crew errors.
But surely with an automated electric car, every event which is considered out of the normal, every running out of volts, every failure properly to identify an object in the road... is noted, analysed, and all use of similar vehicles halted if the event is severe enough until and unless it is fixed?
What? It isn't? Why not?
Partly it's not because car manufacturers in some countries apparently do not have to cooperate with investigating authorities. You may recall
https://www.theregister.com/2020/02/26/tesla_apple_death/
This article is related to a fatal crash of a "self-driving" car in the USA in 2018. It contains the paragraph:
"Members of the National Highway Traffic Safety Administration (NHTSA), questioned during the hearing, said Tesla snubbed requests to describe how its cars were designed to operate under specific conditions, and that Elon Musk's engineers did not intend to take any actions regarding NHTSA’s recommendations in its previous safety reports."
I believe that aircraft manufacturers generally cooperate with the different air-accident investigation authorities.
It's 2021 and commercial aircraft still don't have the ability to weigh themselves? I would fit strain gauges in the landing gear which would not only allow the accurate weighing of the fuelled & loaded plane, but also indicate that weight distribution was acceptable. Counting people is horribly inaccurate.
There is a slight snag in that reasoning as it doesn't take the weight of the fuel into account. The weight of the fuel added to fill up the tanks is accurate enough, but unless the tanks are topped off (rarely, unnecessary weight most of the time), it is pretty hard to know the exact weight of the fuel left over at engines off.
Back of the envelope calculation here: if 150mph wind produces 70535 kg force (max takeoff weight of a 737 according to page 1 of the internet) of equivalent lift and lift force is proportional to airspeed squared, that constant of proportionality is 3.13kg per (mph^2).
So at a modest windspeed of 10mph, that would be 313 kg equivalent of lift. Or about 1.7% of a 737's maximum fuel load.
Not sure if that's a big figure or not. Either way, it doesn't seem to hard to take a fluctuating weight measurement and establish the maximum value, assuming that was when the wind speed dropped to zero by chance. Granted, the flight crew probably need this weight figure way before anybody starts boarding, so they can calculate the fuel load needed. Maybe this is why planes don't weigh themselves, because it won't work operationally?
The gauges would be for take-off weight and therefore for the required minimum take-off thrust, not for fuelling calculations that are done before anybody boards the plane. Fuelling already requires them to carry a substantial buffer so that should be easily able to compensate for headwinds/tailwinds/extended taxi-ing/being diverted, or being stacked for a couple of hours prior to landing.
'"The system programming was not carried out in the UK, and in the country where it was performed the title Miss was used for a child, and Ms for an adult female, hence the error," the report says.'
It's amazing (and deeply troubling) that decisions about important detail of this kind are left to software developers rather than being specified by the client - but it explains a great deal about the growing inadequacy of systems.
I wanted to rant about that, too. It's bad enough that they are deducing a person's age and gender from a part of their name. But, even worse, they are rewriting someone's name to fit their assumptions. Sorry our software won't let women be called Jack - it'll assign a male weight to you - so we'll have to call you Jackie.
And what do the hell do they do about gender neutral honorifics? ("Yes, I did gain my first PhD age 13...")
>Have had the misfortune on a couple of occasions of other's fat spreading over the arm rest into my seat area
I once had someone spill under the armrest from her seat. Luckily it was only a short flight and I could move towards the aisle a bit. I felt very sorry for the guy was trapped in between her and the window though.
In the very early days of public air transport they did indeed weight passengers and luggage in. Now only freight gets actually weighed. The rest is estimated, primarily because (errors of this kind excepted) tolerances are now much wider as planes have greater safety margins.
This is the bit that struck me from TUI's response. Paraphrased - 'We fiddled with some potentially safety critical code which was meant to make things better and wasn't meant to break anything but we couldn't be sure and launched it over the weekend without having anyone on stand-by in case we had broken it'...
---
"On 20 July, 2020, the programmer was making enhancements to the program to improve its performance," the report says. "This should not have stopped the program from working, but as this was a 'fix,' it could not be known for sure. A combination of the [TUI] teams not working over the weekend [to make manual corrections] and the 'online' check-in being open early on Monday 20 July, 24 hours ahead of the flight, meant the incorrectly allocated passenger weights were not corrected."
---
Could some of you more knowledgeable folk confirm or deny that this is a pretty shocking response? Surely you'd do some proper testing on something like this. Christ our compliance team have us run 50 pointless iterations if we make so much as a minor change to a formula in Excel...
As the report says there was never any risk to the flight. 1200kg on a plane of that size is not really much of a difference.
*IF* the plane had been operating at the limits of safety, for example on a very short runway then maybe it could have been an issue, but it wasn't and it's very unlikely that such a flight would ever be taking off from a very short runway in this country.
When taking off the pilots will pre-set things like the thrust they are going to use etc. but that doesn't mean that they just use those presets. Just like you wouldn't think "I'm going to drive up this hill on exactly 50% throttle in third gear" in your car. Yes you have an idea of what gear you will use and how much throttle as you approach the hill, but you wouldn't just keep those settings if you found the car was going too fast or too slow or you forgot that 27 stone uncle Bob was sitting in the back.
The biggest issue with getting weights wrong could be the fuel load, but commercial flights operate with massive safety margins on fuel for several reasons. They don't know if they are going to have to stay in the air for longer than expected, the weather conditions may not be as forecast and they may run into an unexpected headwind, etc.
"As the report says there was never any risk to the flight."
One of the main purposes of the AIIB reports is to communicate lessons learned so others can benefit from them. So an incident which in itself would not have led to an accident can still be highly relevant because in other circumstances the outcome might have been much more serious.
I haven't had time to read this report yet, but the lessons learned are likely to relate to software specification, software testing, management of change, and intercultural communications.
Here's one for the weekend.
Watch uncle Bob Martin's videos, they are amusing BUT they contain very important lessons. The engineer (hacker in this case) should have tests that ran before and after the change showing the effect and allowing consultation on whether that was in fact correct. A quick hack without test in a safety critical system should have EVERYONE involved from the hacker to the CTO in prison. Sloppy practices, lack of training and piss poor behaviour endangered people for zero reason when the job could have been done properly because we know how to do it properly. If this sort of thing is not jumped on by the industry a huge number WILL die and legislation WILL be imposed. Can anyone tell me they still have faith in the computer systems running their cars, after Boeing those in planes or after Tesla self driving vehicles? Hell if we can't play guess the weight reasonably how do we expect to drive a car along a motorway or in a city?
There's a few points to remember here when it comes to weight.
Every plane has a maximum take off weight, which includes the passengers and the cargo. Cargo is easy enough as it's weighed. Passengers are more difficult given the variations between people. If the plane is too heavy, it takes longer for it to reach V1, at which point it could run out of runway and not take off at all. Or, if it does, the centre of gravity won't be correct, stall speeds are lowered (I think), it makes the whole control of the aircraft more difficult. Further, if your plane is heavier than what you think it is, flight levels may not be available to you. Or, you burn more fuel to reach the altitude you've planned. Which leads on to the real crux of the matter for me.
The further aspect though is fuel, which also contributes to weight. If you're flying to a destination, operators will give it enough fuel to get there, plus some for contingency. Ryanair fill it with the bear minimum it can get away with, and has been caught out several times declaring fuel emergencies. If your plane weighs 500kg less than the previous flight, then you can put less fuel in it. It can fly higher, using less fuel, and less fuel means less costs.
The bit about ryanair is nonsense. Minimum fuel levels and reserves are mandated by the air authorities. Declaration of an 'emergency' is also mandatory at the point where you _begin_ using the reserve. In other words, you must land with the reserve intact if you haven't declared an 'emergency'.
Unless you are making your last flight ever, what consequence, beyond a bit more weight from fuel, would it make to carry less fuel? It isn't like it will not get used on the next flight.
In my car (before I bought a Tesla), I always filled the tank to full. Unless you just like stopping at gas stations and pumping gas, filling up the car makes more sense. Theoretically vehicle weight plays into MPG ratings otherwise why would auto makers work to shave ounces and pounds off of a car's weight to the point of leaving out the spare tire (safety much?)
The point is this, on an annualized basis, would you really be able to tell that an airline saved X number of dollars by not putting adequate fuel in their tanks? I would guess no, considering all of the other variables that come into play.
For planes (and rockets): it takes fuel to carry fuel, so from that criterium alone any amount over the absolute minimum is wasteful. Though with planes the airline planners also factor in the cost of fuel at the next airport, as it may turn out cheaper to take some extra outbound, and not have to fill up as much at the other end for the return.
Car fuel efficiency is very strongly influenced by drag (speed, squared) as well as rolling resistance. Not carrying a spare tyre is at least as much a matter of the space it takes than of its weight.
As has been mentioned, it's not just about fuel economy. You want aircraft to be light because they need less runway to take off and to land, they climb faster if they are lighter. Large commercial aircraft can have maximum take off weights > maximum landing weight, so there's a direct trade-off between passengers/luggage and fuel, especially on shorter journeys.
A 737 is burning 3 tons of fuel per hour. So lets imagine a 5 hour flight, 20 tons of fuel needed (with a reserve ) . It has a 40 ton fuel capacity. So you could top the tanks and fly with 20 tons of 'spare' fuel. Thats' the weight of quite a few Benidorm bound Brits (200 or so even at 100ks apiece).
I found a calculation that suggested that by reducing 'spare' fuel by 2% an airline could save 30,000 Euros per plane per year.
Assuming that few TUI flights need more than 50% full tanks, that might imply a difference per plane of 750,000 Euros per plane per year.
Ryanair fill it with the bear minimum it can get away with
I've never flown Ryanair, but all the others I've flown with appear to have had a minimum of zero at the time as there were no bears on board on any of them. At least as far as I could determine, that is, and not counting toy bears or a particular subgroup of gay men.
"The health and safety of our customers and crew is always our primary concern," a TUI spokesperson said.
Clearly their primary concern was getting cheaper software developed offshore and not fully testing it.
Out "thoughts and prayers" are with the non-UK developers and shareholders.
"Well given that Tui aren't UK owned".
TUI Group, (https://en.wikipedia.org/wiki/TUI_Group)
own five airlines:
- TUI Airways
- TUI fly Belgium
- TUI fly Deutschland
- TUI fly Netherlands
- TUI fly Nordic
All those companies have a different history.
TUI Airways has its headquarters in Wigmore House, Luton
It's quite possible that this programming was for TUI Airways only but this article doesn't tell anything about that except that the linked PDF has this to say.
" The operator was the UK associated regional arm of a large European company, with a number of operating bases at major and regional airports within the UK."
To be honest it is time to weigh passengers. Why the hell should pay for an extra couple of pounds of luggage when the heaving fat 35stone monster next to me gets their ticket for the same price as the cute waif over the aisle? It would be especially good to impose a maximum weight for cattle class seats and force these semi mobile tubs of lard into first class. I also doubt that they could realistically evacuate the plane in an emergency with one or more of these grotesque behemoths lumbering up the centre of the plane and destroying the slide by squashing it. There are conditions leading to obesity but largely treatable.
Feeling so much better that "cultural differences" were to blame and not a mistake. When my plane sticks to the ground and grinds into the tower or falls out of the sky through lack of fuel due to a miscalculation I will be able to say 'never mind its just because people are different'
This is so frustrating! People not qualified are writing mission critical software. Boeing had some of the 737 max's software written in India. Ok. Some of the software was written in Bangladesh. Not ok. I'm certainly not implying the bangladeshi engineers are stupid. I am questioning their experience writing mission critical software big time
Let me get this straight then, the Operator has to come up with the mission planning software to work out the payload's mass? And that every operator will have it's own implementation of that? Isnt't this a situation where an international standard would be highly applicable, and one tool instead of hundreds?
The private nature of aviation development has had some high moments. This isn't one of them!
Perhaps more to the point don't airlines have sensors in the suspension that allows detection of weight on wheels for the flight engineers. A little more to those sensors and couldn't they provide reasonable inputs on the weight and weight distribution of the plane?to be honest compared to the tons of the plane, the tons of fuel and tons of cargo I suspect the weight of even the most obese lump of lard is not really significant
That a good specification is actually important and that UK software engineers should be used instead of cheap foreign ones. (Not that given the pauper salaries on offer in the UK there are many willing to take up the trade - next we will be told that there is a skills shortage, but no employer is going to understand why, we are managed by clueless bozos in the UK, people too stupid to be allowed out of bed in the morning)
I couldn't find a reference to the aircraft model on a quick re-scan but let's assume something like an A300 Airbus, which typically seats about 240 PAX and has a gross take-off weight of around 132 tonne. Let's say 200, and if every person on the manifest was an adult female booked in as "Miss" that would be a miscalculation of 200 x (69-35) kg = 6.8t or 5% of gross weight.
Seems a bit over-dramatic to make out this was some kind of huge thing. Bear in mind the actual number of people misrepresented is going to be a small fraction of the total, certainly less than half, so closer to 2%.
From the PDF you find this.
SERIOUS INCIDENT
Aircraft Type and Registration: Boeing 737-8K5, G-TAWG
No & Type of Engines: 2 CFM56-7B27E turbofan engines
Year of Manufacture: 2012 (Serial no: 37266)
Date & Time (UTC): 21 July 2020 at 0500 hrs
Location: Birmingham Airport
Type of Flight: Commercial Air Transport (Passenger)
Persons on Board: Crew - 6 Passengers - 187
Injuries: Crew - None
Passengers - None
Nature of Damage: None
Commander’s Licence: Airline Transport Pilot’s Licence
.......
PS. TUI only has Boeing aircraft right now.
So how about specifications? For software that is considered safety-critical, who left it up to programmer discretion to decide if "Miss" means a child or an adult? I worked on some software recently to automate cryptocurrency trading in a few specific circumstances, risks were relatively low (if it bought some cryptocurrency, it can just be sold back off after all) but I still made sure to go over failure modes and such with them rather than me guessing what they'd want (If the system says a trade failed, it's unlikely but possible it made the trade but didn't return a success message; should I just message the user to tell them what happened, with small risk of missing a lucrative buy?, or message the user and attempt to resubmit the trade, with small risk of a duplicate trade?)
I would expect a COMPLETE list of terms and what weight should be assigned to them (or probably female child and adult wieght, male child and adult wieght, and a list of which term matches to which weight). I would expect the time for the programming team to decide "Miss" may mean a child to be when they see this table, ask whoever wrote the specs "So, doesn't Miss mean child?" "Nope!" "OK then."
The weighing of the plane (via strain gauges in the undercarriage) would be done after the stairs/bridge were detached from the plane, but before pushback. This would give an accurate take off weight unless the wind speed was substantial. It would include the plane, fuel, passengers, luggage, etc. in fact everything except the bits of the undercarriage below the strain gauges, and those bits don't vary in weight much so can be added as a fixed value.
This weight would then be used to calculate the take-off thrust needed rather than using software to "take it's best guess" on what that weight might be. I say guess because it is based on average passenger weights and estimated carry on luggage weight which are going to vary massively.
I'm a big bloke (6'4" and very heavy) and would quite happily pay more for my flight if the airline would provide me with a seat I can f***ing fit in. At the moment my choice is economy at say £200 where the seat is too short for my height and too narrow for my wide shoulders and fat belly/backside, or Business class at £1000 where the seat is just barely big enough. I'd quite happily pay even £400 if I got twice the space of an economy seat, but that seems to not be an option. I am also not allowed to book the seat beside me as an empty seat, not that that helps much as my legs are still being mangled by the a***hole in front insisting that he just HAS to recline his seat. The upshot is, I avoid flying like the plague as it's a s**t experience all around.
" we corrected a fault identified in our IT system"
Sorry but the fault was not in the IT system, that worked perfectly as designed. The fault lay in the project management of the software delivery, any QA system that was in place and the oversite of the software development that should have stipulated all of this up front, checked it was implemented correctly and and tested using real world example data.
As always it is so easy to blame IT as the problem when it is lazy ineffectual humans the screw things up. Computers are totally obediant morons that will do exactly what you tell them to do, if they don't do what you want them to do that is your fault not theirs, you (should) know their world, they do not know yours. (you can also say much the same for developers from other countries)
According to this saga, the worst mistake was made during a correction when they already knew it was iffy. They have (you would hope) years and years of archived manifests that they could have compared old vs new software on (regression testing). Run a few thousand of those through old and new, check results within a margin of error. Not hard, just usual good practice.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
