I promise that the last 15 years the IT dept has done nothing except warn higher ups of an impending failure of the system, and the higher ups likely said 'We'll think about it' while stretching the backbone of their company on box fans pointed at server racks in rooms where the AC failed 5 years ago. Always prepare for the worst, because the worst WILL happen given enough time, and however much it costs to prepare, I promise it'll be cheaper than waiting for the worst to happen.
Yeah, at this point I'm almost convinced politicians won't do anything until the system actually collapses. So whenever that happens it'll be a year or two of pain for whoever needs to fly somewhere, as the industry scrambles to rebuild their entire infrastructure from scratch. 😛
The trouble is the IT department generates no income but has big capitals and overheads. Shave some of that and hello promotion! I've been running the IT department at my small company for the past 11 years and the only message I ever get from the CEO is COST COST COST!!!!!
Airline consultant here: this is INSANE. The fact that you included information that even I didn't know is beyond amazing. I only today realized how much research goes into these videos like...insane, even the internal tools wow. Great great job!!!
don't forget the passengers exploit ATPCO through usage of VPNs and get a fare for just 310 dollar for a long haul flight from let's say Dallas Fort Worth DFW to Amsterdam, without VPN and booking from the netherlands it's 1200euro's but connect to a VPN in Mexico it's just 400 and with a discount voucher you get insanely low prices, and... say amsterdam can erupt into chaos it get cancelled but not necessarily due to weather but due to a technical fault caused by icing during departure from amsterdam because they didn't de-ice a really frozen Flap pod. then it classifies as a technical problem. flaps problems and one side of te flaps working good the other not can result in a deadly crash. because of 5cm of snow but freezing -20 overnight in march..... so you get to fly on an airfrance flight on an A350 to paris with a free TGV ticket from CDG to Amsterdam, the fuel economy being way better then their A330-200's of KLM you get to that 310mark and choose your seeting for 10euro's extra. and the passenger who has to wait for the airfrance flight to depart 3 hours later, but upon arriving in CDG just hop on a TGV and transfer to a thalys at lille'europe or in brussels On a amsterdam express which will be operated by highspeed trains as well (the ICNG), the eurostar coming FROM london doesn't require lengthy passport checks. a thallys must be reserved and i dont know if they run from CDG directly to amsterdam let's just say a TGV + ICNG is cheapest for the airline. and why not brussels? well brussels airport is more inland and a flight to there would deviate much further actually causing more problem and delays Part of it is a true story... i experienced. especially the crazy price differences and refreshing pages cache's and virtual machines with VPN's to book tickets, amazing. and yeah rerouted 500south f your destination with a highspeed train a thalys from gare du nord would take 2hoursand 40 minutes with more legroom and a awesome 300km/h experience, the only slowdown occurs between brusselsand antwerp were they they do 160km/h but can do 200 with ERMTS enables but not all trains i belgium and the Netherlands have ECTS that is used with incab signalling. from antwerp to amsterdam it's a 300km/h ride again with a stop in rotterdam. and if you live closer to rotterdam than amsterdamm or even closer to antwerp like TErneuzen or near or in breda in antwerp or brussels you take the brussels express and in HSL zuid in the netherlands thers a point were trains can exit high speedline and get to breda very quickly..... it's just amazing parnetership with rail AirfranceKLM SNCF SNCB DB and NS that allows that option as 5CM of snow is not gonna stop a train
@Teamgeist no way to verify without spending just as much effort... and no effective reason to doubt, in my case. The channel seems pretty legit, this does not strongly affect me, so I can just choose to trust here, without much concern.
The fact that literally 40 years after taking my first flight I still can't purchase an airline ticket using my full and correct legal last name because it contains a hyphen is mind-blowing to me. That really says everything you need to know about the airline industry's IT systems.
@@Evangelionism Can confirm. Can never figure out whether they want me to leave out the hyphen altogether, replace it with a space... and oddly, this isn't on most airlines' FAQ pages online, despite hyphenated last names not being THAT uncommon.
@@inspiratron That s interesting and pretty ridiculous man. My first name is double barreled (although I usually spell it in a way where I don't include the hyphen, sometimes even for legal purposes) and my family names are double barreled. I haven't had any issue traveling before but buying things online using my bank card can be a bxtch sometimes! Couldn't buy myself Discord nitro using my card cause the hyphenation of my family name made it hard to tell whether to use a space or a dash (neither worked).
As an IT professional, corps VERY rarely replace or upgrade hardware until it literally is about to burst or already burst. Its VERY frustrating. Ive been used as a scapegoat for this shit before too.
And trying to explain that the cost of an eventual meltdown is going to be way more painful that an organized, preplanned upgrade just goes in one ear, out the other...
Me too dude. The Bean counters simply don't understand that the cost to replace obsolete ticking time bombs in the environment is nothing compared to the monstrosity of a headache and of a cost that it will be to try to get a company back up and running if a piece of the critical infrastructure goes kaput. Now you're paying all kinds of emergency fees from consultants and emergency rush jobs from other places to get your hardware faster or whatever the case might be, all while the brass are screaming at you to get it back up and running
"Low-cost carrier" is a business model description (like network carriers, charter airlines, and others), not necessarily a price observation. Lower turn-around times, uniformity in aircraft types and seating classes, direct instead of hub-and-spoke flights, etc. are all intended to reduce costs, but don't guarantee drastically cheaper prices than other airlines (especially since these days, many non-LCC's are still influenced by some of the LCC philosophy)
As someone who works in IT, it is completely unsurprising that this is the case. You'd be appalled by just how many vitally important things run on highly outdated tech. This is usually because the cost, manpower, and time required to replace it is often absolutely monumental. So company bigwigs stick with "well if it ain't broke, don't fix it." Even the pentagon still has critical systems that run on windows 95 and 98. Not saying it's right or the best way of doing things, but there's usually more to the story than just pure money or greed in these cases. Nothing in this field is ever as simple as just pressing a few buttons on a keyboard or unplugging a device and plugging a new one in.
Yeah... We work on systems running XP but when I worked on IT airline systems that was a whole new level of ancient. OS2 , token ring, and whole airlines running 56k links
Sometimes updating to the newest software can cause more issues short term than the old software was. And in areas like engineering, software like cad used to be a pay once and own it forever. Now its exclusively a subscription that costs as much as the original own it forever price so no one swaps out the old software going as far as to setup virtual machines to run it.
I work for a software consulting company. Basically my job is to write custom software for businesses. So naturally, I deal with technical debt on a daily basis. It can be very frustrating at times because we are not in control of whether or not the hulking old system is replaced or improved. That's in the hands of our clients. We can always advise, but never command. However, when we _are_ actually contracted to replace or improve old systems, my job suddenly becomes amazing. The level of impact I can have is staggering. My work can affect people all over my state, or even all over my country. I love doing this kind of work, but I have to be asked to do it, and given the money and time to do it right. If you're out there and you're in charge, take a good, hard look at your systems and ask whether they could be considered modern and efficient. Ask what the cost of failure is. Then do something about it, before the whole thing caves in.
I had a new client 6 years ago that had been plagued with constant minor failures and quarterly major failures. After evaluating the entirety of their infrastructure, procedures, and software I recommend a complete overhaul. Everything including the wiring was replaced. New software was deployed, and new procedures were written. It’s been 5 years since everything was completed, I only see them annually now to check in. No issues are ever reported. When you do it right and spend both the time and the money, things just work.
This is a wonderful anecdote, and I also think an example of how important Systems Design is of the most critical things when creating long-term solutions. I also have a similar responsibility with a small data team and think about how doing things with effort can have a positive impact not only on the company but also thousands of customers is rewarding and motivating.
As an ex-travel industry employee (thanks, Covid), seeing that GDS screen gave me some PTSD. Travel agents call it 'flightmares' - stress thinking about doing something wrong in the GDS system (which can result in expensive fines from the airlines if you break booking rules accidentally that the system allows) or having to do some major re-booking. I don't miss flight bookings at all!
As some of the of the Airlines work to upgrade their outward facing system the GDS starts to look more like a bunch of json data. I guess it explains why I've heard it called the "database."
This is crazy to think about. The only reason I re-entered aviation is because of covid. So many commercial airliners need to be stored because flights are still down, which is the reason I even have a job right now (I'm a mechanic at a long-term storage facility)
i heard of the problem of "screen fatigue" from military enthusiast circles its mostly about how long a radar operator could stare into a screen until his eyes feel tired and begin to affect his performance, which is crucial in detecting enemy planes, missiles and stuffs seems like its not really unique to military now
Seems like a perfect work for AI. (not joking, Ive worked some administrative and managment jobs that no one wants to do and everyone would probably highfive if there was an AI to take over, letting us focus on more "humane" tasks)
Thanks for your proper voiceover with natural pauses. So sick of lazy UA-camrs who splice a thousand edits together, chopping out all pauses and creating a never-ending barrage of words where you don't have a chance to think or absorb the info.
I finally retired from IT after 40+ years in it. We would beg and plead with management to spend more money on staffing, upgrades, etc., all to no avail. Then, when something finally did break, it was, somehow, all our fault. Of course, that never stopped them from spending money to implement the latest fad in software development or whatever and trying to shove it down our throats, thus making it even more difficult to do our jobs. Feh!
I'm in IT and a pilot. The cost of upgrading the systems would be horrendous. The amount of FAA and federal red tape and testing would be insane. It would literally have to be all the major Alliances getting together and crowd funding a standard system its that bad. Also, keep in mind Airlines aren't about new technology, they are about stable, and supportable.
@@ooo_Kim_Chi_ooo I don't think you get the essence of comment above - it's not about new technologies or upgrading everything at once - it's about continuos improvement, which is impossible when managers have mindset of "do not touch if it works". Simple as that. Yes, I can agree that system so safety-critical on which many other system depend has a lot of inertia, but that's only argument against revolution, not against evolution. And last, but not least: the amount of money developers are begging for is magnitude lower than money spend on investments in other branches of this industry.
@@proosee Its just not possible though in the airline space. I mean people easily forget that in the 80's planes were literally falling out of the sky and when Airbus introduced all their wizardry it didn't go off without a hitch. The airline industry has been bit hard by new technology whether it be computers or hardware. They are extremely slow to move on anything and approach more of an it's working perfectly now don't touch it until we need to approach. When you are dealing with lives and money at such a scale their willingness to be up to date is lessened by the fact that they really have no really reason too other than its old and shitty to use.
@@ooo_Kim_Chi_ooo still: it's not "willingness to be up to date" - literally entire video is about weak spots of this system. I beginning to doubt you ever integrated with legacy system, because if you say "shitty to use" is the most significant problem then you are doing those kind of integrations wrong - you can easily create some intermediate layer to isolate from it and never thing about it again. The case here is that this system is unreliable, slow and faulty - and yes, there are new technologies that solve those problems, but word "new" is not the most important here.
I worked for about 5 years with a travel agency, and got really familiar with Sabre GDS. The tool itself was quite marvelous as a new agent, but as the years went on I started to notice the huge amount of limitations it had. For example, the entirety of a corporate account for one of the biggest oil companies in the world could be messed up quite easily by one single agent unticking a box while looking up a PNR. Happened many times, and I swear there was never a change on the GDS to avoid delaying thousands of passengers from booking their flights.
Hmm, what box is that? Granted, I don't use the graphical interface (I've been on this GDS for 30 years, don't need it). Are you referring to a command that allows the editing of the 1st level star?
When I went to college in the late eighties, one of our profs had experience in the airline industry and helped with Sabre. Fast forward to 30 years later, we are still using the software he worked on. How sad is that.
US Airways had updated their reservation system and the agents liked it before America West bought them. America West brought Sabre back, and then they stuck with Sabre through the merger with American...yay!
I'm a tech for a company that provides IT support to airports all over the world. Sabre, IMAGE, and whatever the heck Delta Airlines uses are incredibly dated and clunky. They look like an app running in Windows 3.1
My mom worked for Air Canada in reservations and bookings in the late 60's and early 70's. They had one of the most up to date systems at the time. It sounds like not much has changed since then, and that she could literally do the same job 50 years later with almost no additional training. (She still remembers most of the 3 letter airport codes too!)
@@wta1518 why do you say so? Personally I like 3 letter codes just because I'm so used to it, and they usually have some resemblance to the location of the airport.
My father worked in ticketing many years until retirement. He used to work with an Amadeus terminal and type the most complicated reservations on the command line by hand. He then was unable to work with a windows computer or use the mouse 😂 This brought good memories, thanks for the video.
It's not a simple case of management not listening. There's no incentive to listen as the video says. You upgrading mostly favours the competition then you. American Govt or international airline organizations could easily fix this by mandating the new solution within a year for all airlines. But they aren't doing it.
This isn't just an issue with Airlines, as someone who's worked in enterprise IT for Large companies and Governments for a few decades now, everything may look nice, but almost all system backends are running on either an ancient system, or ported over from such (The old software just moved to newer computing systems with a new-ish user interface, but the backend is still talking in what most would call "dead" protocols and database languages. Cobol, Fortran, etc.) Same goes for those "fancy" systems that healthcare use, it's all based on a 1960's protocol (M, or MUMPS) The problem is they become so used, that nobody wants to be the person that changes it out in fear that it'll break things more, and or deal with the downtime involved. On top of that most "new" systems are being developed by "Cloud" software companies, rather than software you can own/run, customers are being "locked" into a company that can... once under contract, be able to hold your data for ransom and charge whatever they want. (This is the main reason behind the "cloud", it's all subscription "locking" in customers, not about saving money or reducing support needs or cost).
i agree with all of this but especially your last point. between all the different saas/cloud companies that either go out of business, get bought out & slowly scrapped for parts by a larger company, or completely overhaul their product/company focus out of nowhere-- making the decision nowadays to try to upgrade legacy services is a real rock and a hard place. and that's not even to mention the challenges around sourcing new hardware these days (ironically because you'd be competing against those same cloud companies massively scaling their data centres)
Can't companies create a Linux system or they should look at India. There are lot of firms who will make non cloud software that will give ownership to company
The difficulty with the legacy code in Cobol is the lack of tools we became used to in the last decade. Automatic deployment and thorough automatic end to end testing are a hell of a drug. And even more expensive when to be introduced after the fact to a system that is about six decades old and full of tightly interwoven ideosyncracies noone knows fully or may even be in a position to start to understand. Re-implementation is going to be expensive, but for what gain?
@@sorryi6685 yeah, Linux is great for servers, but no company which isn't IT focused and wants to be successful will touch Linux. With Linux, you as a company have to operate and maintain all of it, while you as a company do not have the expertise and not the resources to stay up to date. The same way a production company doesn't run their own haulage subsidiary.
I agree with the first part, but not with the cloud/saas. Does a production company operate their own haulage subsidiary or have inhouse Architects? No? Why not? Bc it's not your business model, so you should only keep the minimum inhouse infrastructure, the rest you try to outsource if possible. The bigger problems are company mentalities, who think that the IT system has to fit all processes, instead of finding a middle ground - pick the software which fits your processes best, and adapt your real world processes as much as possible to the IT landscape. Cloud means you outsource cyber security (only a few companies can afford an actual, decent system inhouse) as much as possible, you don't care about software updates which can stop your production, you minimize the hassle of finding good IT guys (especially if your company is in the country side and a bad IT guy is always worse then none), you need less expensive hardware to run demanding software, lower capital cost and better scaling (classical example, Photoshop - one time licence is ridiculously expensive, a monthly payment plan which flexible in-/decrease is create. Or how big is your server supposed to be? Doesn't matter, take the small one and book additional capacities when needed instead of overscaling your IT hardware landscape) Has cloud downsides and do you need external consultants to support your decision (not making it for you) which one to pick? Obviously. But these negatives dwarf by the alternative - keeping everything inhouse.
If they needed to be bailed out then it should be a publicly held company that is sponsored by the government and run by the people, not a corporation, since the government essentially bought it.
@@t.j.vellinga6225 Unfortunately, many of these guys are "too big to fail." Boeing for example, they have been a shit company that ignores safety protocols and all about cutting corners to save cost, yet they keep getting gov contracts to keep themselves afloat. It's because Boeing is one of the largest employers in the US, and if they went belly up, tens of thousands of Americans will lose their jobs. Boeing is holding the American people hostage to force bailouts and contracts from the gov. It's the same thing with many of these big corporations. Their argument is always "If we crashed, we will take down tends of thousands of jobs with us and it will hurt the US economy more than just giving us a few billion dollars to get bailed out."
In the US the airline industry is already subsidized by the government and also has contracts to fly US mail. Any more money given to them will go straight into the airline executives' pockets.
Google Flights doesn't do a GDS lookup. It buys availability data directly from OAG and consumes AVS messages from airlines directly and use a much smarter data structure internally to quickly evaluate farings.
Yea, I found it weird that the way it was described as being bad is that "behind the pretty UI is a system that joins data from multiple APIs" like yea... that's the entire IT sector right there lol
Iirc Google bought ITA Software, which powered this kind of search for a while and even supported some small airlines, maybe just one Caribbean one. And they got a ton of flack about anti trust from a lot of air travel businesses for basically side stepping the GDS.
@@getellied I was thinking the same thing, but one of the key points was that those APIs haven't really changed in any meaningful way. I'm guessing GDS is mainframe based, which it's pretty hard to find developers for now, and architecturally change would be incredibly risky. Typical enterprise risk avoidance, where they avoid risk so hard that they end up in more risk in the long run.
I went to flight school through college and got partially through , then dropped out to work IT haha. Loving working as a field service tech right now. The college program (Eastern Michigan University) was horrible; I was told something like 48/60 of students in my year ended up switching out of the flight science path and got av management degrees or MBAs. We had to take multiple semesters of aviation-oriented classes that were almost entirely useless (stuff like weather science, adjacent to aviation but not applicable in a pilot's career) It's hilarious to me how important the aviation industry is to society in general, yet the barrier to entry is so artificially raised by school institutions and the FAA. College programs are the most approachable way for kids to get into aviation beyond those whose daddies can pay out of pocket for training. We'll continue to have a pilot shortage until young pilots aren't made to jump through a thousand hoops just to get their degree on the academic side *in addition to* flight school It's ridiculous to me that stuff like VORs is something you're still required to spend time and money on and understand, yet they don't even exist anymore really and were mostly obsolete since the late 80s. But I digress lol
As an aircraft mechanic, I knew our it systems were fucked in regards to parts procurement, but never imagined other sectors of aviation to ba also fucked to such an extenr
Then you haven't learned sh*t... if you are a pilot you should have picked up on the fact that this dude doesn't know enough to know that he doesn't have a clue.
I was an IT professional at a bank. It frightens me that some of the processes that the bank relys on are still running in Cobol. That code is as old as I am and I'm retired.
@@knoahbody69 The only Cobol code I ever wrote was as a part of a course while I was in university. Even then, Cobol was considered to be old, obsolete and strange. No, I did not write any Cobol for any client ever.
I develop IT infrastructure topology to companies. And, o boy, what I see in the backbones of IT infrastructure is insane sometimes. Many companies don't have a budget constraint in infrastructure, specially when they are the basis to the business or when they have a cultural bias towards innovation (like tech companies and startups in general). But some companies are properly insane. I recently run a metrics analysis in a BANK, and what came back was... wow. People running HR, PD and salaries through Excel sheets, with no backup... IT personal making adm backup into their personal Google Drive accounts... Memory and latency beyond critical performance with 7 years old hardware with no support or spare parts... Crazy after crazy
Honestly though, I see this more as a problem in the IT world than non-IT. Upgrading an old vehicle to a newer one is much more straightforward than upgrading to a new software solution. Software never works as intended or as expected and workers find ways around that, but not even good training can overcome the experience with the old system. It's like giving out a new vehicle with the steering reversed. Sure, on paper, just steer the other direction. But of course everyone would crash.
"If builders built buildings the way programmers wrote programs, then the first woodpecker that came along would destroy civilization." -- Weinberg's 2nd law
I worked in small business as part of an MSP. So many HIPAA violations. No backups. 15 year old virus ridden servers. It was disgusting. Never use your credit card at gas stations, Dentist offices, lawyer offices, even hospitals.
this is hilarious because this is exactly my experience... when your office workspace still uses a more primitive version of a spreadsheet you know whats up and I have no incentive to advise the change so... why bother... -_- infact it got so bad that I want to try ask airlines how the heck they still get parts for such PCs cause finding them is getting rarer in my part of the world in short that yes I want to adapt what airlines use for the house... its THAT BAD tl;dr the office still uses DOS applications to run... and I get laughed at posting comments at LTT if they have answers... well LTT community being LTT community you get the idea
hate to break it to you, but that is common with banks. They don't have IT as part of the base mandatory infrastructure because technically one person with a spreadsheet can run a single bank...just not process more than a few customers per shift ;)
This "Good Enough" mentality with accountability only to the share holders is honestly one of the most pressing problems we have today, period. Not just in the airline industry, but in industry altogether. It's a primary driving factor in the income gap eroding the functional middle class, driving a portion of global inflation and encouraging corporate corruption on a massive level.
Yes! Everyone has bought into the "we can fix all our problems by cutting red tape" mantra that politicians use to pretend their budgets are going to balance because of those magic red tape savings, and everyone seems to have forgotten that this is PRECISELY what regulation is supposed to be for. Making sure that the thing the free market is optimising for is actually the outcome we want, because they will default to whatever happens to be in shareholder interests and sometimes this is in the public interest and sometimes it isn't. It is just much harder to actually do the work, think it through, and implement something rather than just make noises about evil "red tape". (And if we are honest, how many voters are going to recognise that work as being valuable?) If the FAA funded a new fast, modern, extensible system and then denied licenses/permits to every airline that didn't update to support it, airlines would fall over themselves to implement it and not lose the revenue. Problem solved.
as someone who works in IT, you'd be surprised how much critical infrastructure runs on ancient, fragile code written many decades ago but hidden through a nice, modern interface
Yeah, I work as a building automation specialist and it's crazy all my clients thar still do that. The only industry that I contracted with that never was like that was pharmaceuticals.
I don't have concrete evidence, but I suspect hotel and car rental companies do this by tracking browser cookies and IP addresses. Sometimes I have secured better rates when using a VPN via another state or country. And don't always believe the "Last room left at this rate" flags.
@@simongchadwick Fun trick for hotel booking. Check your booking price, and if its to high for you, reset your browsers cookies and go back to the webpage. (Do not reload, input the original link again.) On certain booking sites, this is enough to have the algorithm generate a new price.
As an airline pilot, I can confirm that this is all true. Our scheduling system still runs on Windows XP. An early version of it. I’m not kidding. Also, regarding the experienced rebooking agents part of this. Also very true. Just a few months ago as I was about to fly a flight home where my sister was a passenger, I realized it was going to cancel. 2 hours before it canceled in LGA, I knew there was a flight home out of JFK at a slightly later time and was able to get my sister on the train to JFK so she made it home. Just knew that because when you do it day in and day out, you eventually figure out what goes where and when.
@@hassanabdulaziz7275 Worldwide. Ask any air crew and they've used a horribly outdated software at least once, probably still does. There's probably only one good EFB software in the market right now and even that is not great
The “reverse tragedy of the commons” you described is basically the prisoner’s dilemma, mutual cooperation is beneficial on aggregate but individual non-cooperation is locally better
I think a better word for the "reverse tragedy of the commons" is capitalism. Why would individual non-cooperation be locally better? Competition is a disease preached to us by the mutually cooperating owner class.
@@CC3GROUNDZERO Go starve in Venezuela and then tell me how well the alternative works. Oh wait, you wouldn't be able to because even if you survived you aren't allowed to complain.
Great video. As a former IT professional I've used the economic uncertainty of today to try out some other fields besides tech. It's crazy how you can tell management that a planned upgrade/fix will cost so much less than if the system explodes, but they never listen. Then of course when things do explode, they try to throw you under the bus. I've always made sure to save emails of this happening, and it's saved my ass before. I hope things change but I sadly doubt they will.
but why are IT systems not designed to be easily upgraded or transferred to new systems. seems to me that should be the main requirement but it's always so complicated and takes great resources.
@ron black The best way I can describe it is to imagine it like changing the gauge of rail track. A larger gauge functions very similarly in concept to a smaller gauge, but you still have to replace the rail itself and the wheels on train cars. Then, there'd be a transition to the new system where a loss of productivity occurs until everything adjusts to the new system. It's pricey in the short-term but not as pricey in the long-term as overloading a smaller system until it fails. Short-term profits in capitalism still prefer to keep the smaller gauge and hope for the best. In IT terms, businesses are using old code/hardware and hoping that it holds instead of transitioning to new systems with more capabilities. Code would need to be optimized for multi-core, hyperthreading capable processors, and parameters would need to be able to accommodate the massive amounts of data an upgrade could provide. Things get complicated quick when you're dealing with trillions of ones and zeroes that need to be processed then sent somewhere else
TBF a lot of IT professionals do also want to upgrade things that are working fine and not about to explode. We had to pull teeth to get an expensive ($50k) bit of network testing automation equipment, but we could have still done the tests with less automation, it would just be more work for us. The equipment was a big switch that let us set up test scenarios automatically instead of having to hard-wire them.
I'm a computer scientist from germany and I knew the guy who came up with a similar system for German Rail. Whenever he heard about a delay of a train he stood there for a while frozen calculating the delays impact for the whole network.
And btw the key to solving NP-hard problems is using heuristics. Otherwise, your system will only be good up to an engineered limit and will reliably fail beyond that.
Understand too that Europe's national rail operators often operate different & incompatible ticketing systems, & people desire them to be more standardised & integrated like GDS is for aviation, so as to facilitate international rail travel
My friend is studying aviation (basically pilot degree) and they asked me if I wanted to see their flight computer. As someone studying computer science, I find computers very interesting. They pull out a paper with a paper wheel pinned on and hand it to me. I was so confused.
Different thing all together. Pilots carry E6B in case of electrical failure, where an analog computer may be all they have to do certain calculations. Technically it’s a flight computer, at least that’s what it’s called, but it’s more like a flight calculator.
I trust an E6B more than I trust anything else. It's what I learned on. In the event of a total electrical failure you will be returning to your E6B, lap board, and pencil. Total reliability.
As a computer science student, you will eventually learn that "having the latest tech" is not only not achievable, but also dangerous. Robust and failsafe systems are preferred for critical areas, and in case of electrical failure pen and paper are king
I used to work for an airline using Amadeus. Unfortunately they started going with a graphical skin on top of Amadeus and forbidding us to use the command prompts we were using. That meant that all that changed was that now we were only allowed to click on buttons, which added significant loading time per click. Worse, the system didn't actually change, it just limited what the agent was allowed to do. So once it crashed or broke (which happened all the time) all you could do was wait until the system sorted itself out. Glad I left that job, so much frustration for nothing
I work at a company selling a GUI rivaling the one from Amadeus and some customers are so happy to switch to ours because it's so much more simple and responsive
@@samconstantinou2335 if you're allowed to just use amadeus you'll be fine haha. Very clunky but once I got used to all the command prompts I was so so fast
Manager's response: "solveable? I don't think so, there's not ticket and no budget for that - go working on something useful like new UI design for customers so it will look nice, but be more confusing for the user".
@@proosee of course their is money, it's just a matter of if they want to. All they need to do is have these billion dollar companies all put a bit of money on the side to fund a new or existing industry group to put in the work. Lots of industries have some done similar things. for example the banking industry has SWIFT.
I worked at a major international airline during a period where it realised its' old 1960s reservations system was no longer fit for purpose. The options were a) build a new modern system from scratch b) apply duck tape and bolt some clever little fix programs on top of the existing system. The airline chose b) because senior management was (probably rightly) terrified by the risks associated with a) - development effort, time and cost / amount of testing required / smooth cutover / integration with other legacy systems / PR damage due to inevitable problems. When your business is almost completely dependent on IT systems, reluctance to change is understandable.
The failure rate of replacing the old airline systems is astronomical. So they have instead focused on building additional functionality outside of the old GDS systems to supplement the old limitations. But it is all very complex and spending money is no guarantee of success.
Yeah, but then you do this for 60 years and no one is now able to even just duck tape the crumbling mess, much less reconstruct it in modern technology (especially since the documentation is lost to time if it even existed in the first place)
Legacy stacking will NEVER go away as established companies/ industries such as the airlines and telcos aren't willing to spend Billions on developing and implementing new and improved systems. Biggest problem is the further you kick the can down the road, the more complex and expensive it gets. Sabre was developed in the 1960's and is the main reservation system still used today.
European passenger rail companies looking at the situation, after having done the exact same thing and now refusing to go back to interoperability in booking systems: finally, a worthy opponent!
weren't they given an ultimatum by the EU though? Where they have to collectively present a plan on an interoperable booking system, otherwise the EU will make their own and the companies have to use that one. EDIT: granted, that would still mean it is at least a while before it is used.
@@zephyros256 they were given an ultimatum, but I sure haven't heard anything of them even bothering to do anything about it, lol (if anything, some were actually lobbying for the EU to abandon the idea). And if the EU decides to make something of their own, we may well never see it in our lifetimes.
You could have formatted this comment in a non zoomer ticktock algorithm way. "Me typing out an entire comment with context, sentences and punctuation: Im gonna ruin this mans whole carrer!"
The really interesting part is you can kind of apply this stuff to basically any industry that was big enough to get computers going in the early days. Banking, Hotels, Medical, Airlines, Power, Shipping, ect.
This is an exceptional video. Growing up, both of my parents were flight attendants, and helping them navigate the command line interfaces of their respective airlines is what got me into computer science. One other tidbit - their bidding process was absolutely INSANE by modern standards. When flight attendants have the appropriate level of seniority (seniority is what EVERYTHING is based on in the industry), they have the ability to pick and choose which flights they want to work. Back in the day, this was done by distributing a gigantic packet to every flight attendant which contained every flight itinerary for that airline for the entire month. When I say gigantic, I mean it. The bidding package was often 600+ double-sided pages, in tiny font, with almost no white space. The amount of paper they went through was absolutely ludicrous.
My dad's best friend when I was a kid was a United pilot and his wife was a flight attendant. I remember them comparing notes because my dad was a truck driver and he bid runs as well. The airline stuff was absolutely crazy. The upside of it was when my dad was gone on the truck and Dale was in town I used to get to go to Stapleton and hang out with him once in a while.
I used to be a travel agent for a large travel agency based in NYC and had to create itineraries and fares by hand using commands in Sabre to generate tickets. I even used validation plates and had to hand write them as well depending on the air carrier.
We have those as well when I was with KLM. Whenever I use that validating machine to issue an MCO, the passengers ar the counter are always startled with the loud smashing sound it makes. 😂
@@vishalsyoutube A validation plate is a metal embossed plate with the airline name and logo along with their three-digit IATA issued accounting code. This allowed you to validate an airline ticket on that airline or one in which they had an interline agreement with.
The other issue with replacing and updating the GDS is that there are other industries that also rely on it. Hospitality, car rentals, some entertainment/tourist stuff, and even some bussing and cruise ship systems. There have been a couple different pushes to update the GDS by big players over the last 20 years and it never moved past the idea stage. One of my previous jobs I worked at one of the main Hospitality companies, we updated and rebuilt our systems every 10-15 years but they still hooked into and used the underlying 1960's GDS.
I can tell you from experience in an unnamed payroll company that changing from a command line system to a graphical one for the ticket agents will NOT make things smoother, faster or easier. When you know how to run your command line with the amount of information needed you can fly through the process. Make that a point and click and the delays between even just screen refreshes will turn a 5 minute process into 20 minutes.
True, but there's still no reason why you couldn't invest in improving the system - making the process of getting an re-booking more flexible, scheduling including crew information, making it harder to do a mistake - or make the mistake less hard to fix afterwards. From my experience I want to echo what many here are saying: IT might be expensive, but the one of the things even more expensive is to wait until it crashes and everything burns. How much did SW have to pay for their Christmas chaos? You can probably save up to 90% of that by improving your system before it crashes in the worst possible situation. Down time is bad, but down time during Christmas is worse. But as the IT guy, you don't get to hire the people needed to fix the system, you're just there to keep it running, day after day, while yourself most clearly seeing these things will - one day - crash and burn an you're like Cassandra forced to have no one believe you, or no one care.
A lot of the improvement comes in training new employees. Text systems are deceptively fast for experienced employees, but painfully slow to adapt for the new.
@@Steven_Olson that's why TUIs (text user interfaces), which are a hybrid of CLI and graphical, are amazing. Powerusers can rip through using hotkeys while noobs can point and click their way through menus.
The decrepit IT isn't just on the airline industry. I work on the UK rail system, and we similarly have a back end system for train schedules (called TRUST) which is the backbone of the timetable information supplied to passengers and the public. There are other systems too for things like crew scheduling and rolling stock planning, but these often don't talking to each other, so when a major route gets closed for an unplanned period of time due to an infrastructure issue, or emergency, that's why it all falls apart so dramatically, affecting services often far away from the area where the problem is.
"There are other systems too for things like crew scheduling and rolling stock planning, but these often don't talking to each other, so when a major route gets closed for an unplanned period of time due to an infrastructure issue, or emergency, that's why it all falls apart so dramatically, affecting services often far away from the area where the problem is." Indeed. I think the video poster is trying to imply that we should try to replace aviation with public transit trains and that would solve it, but public transit does the same kind of stuff. And it doesn't have the speed that we need in North America and Australia. It can't take 20 hours to go from LA to NYC. Five or six hours is already too long. Good news is that Lockheed and NASA have developed a fuselage shape that lets you break the sound barrier with a barely audible sonic boom, so hopefully we'll get supersonic over land again....so it will be more like 2 or 3 hours to go from LA to NYC. 🙂
As an operations/aerospace engineer inside Boeing, I am always impressed by how concise and entertaining your videos are! This one take the cake and yet its just the tip of the iceberg in terms of the constraining factors that GDS has within the part 121 world.
What giant industry or corporation these days *doesn't* settle for good enough? I've seen it happen over and over. Businesses have to be good to get big, but once they get big, they stop trying to be good and start trying to stay big, and being good and staying big are sadly different directions for most businesses.
IBM, Goggle, Facebook, PayPal, airlines, manufacturing, railroads, trucking... list goes on and on. It's like the old saw of running from a bear... I don't have to be the best in the industry. I just need to be marginally better than you in some manner.
To be fair to airlines, they are alot more regulated, and most of their stuff has to first get approved by the FAA before they can use it, and that's expensive and slow.
7:54 - this sure brings back a lot of memories accessing my former airline’s reservation system to check bookings and flight availability. Had to have a cheat sheet to remember all the codes needed! As for Southwest not allowing booking via GDS - it’s a legacy of the early days of the GDS. Back then, those GDS were owned by the airlines (for example American Airlines owned Sabre, while United owns Apollo which was merged into Amadeus). These airlines, in a way of trying to squeeze their competitors, chose to charge other airlines for every booking made on their GDS. Even worse, some of the airlines, namely American, dicked with the algorithm of the GDS so that their flights get first display on the GDS because they found that travel agencies back then would often offer customers the first flight offered because they were used with the fastest flight being first in their OAG routing book. In one instance, they dicked the system so that flights from New York Air, a competitor, was listed last in the system. Why? Because they dared to challenge American on the New York-Detroit flight. All that led to Southwest to say, “F*** you, we just won’t sell on your asswipe system” And that’s why they’re not on a GDS today.
Ι am working for the reservations department of an airline and although relatively newbie and a young person, I heavily dislike the web format of our GDS (Amadeus ARD). I was trained in it, but I surely prefer the old-school cryptic commands. In reality, it is a lot simpler and faster to work this way. Instead of clicking "Reprice" then choose the passengers, hit "Confirm" and then "Update Fares", I just write FXF and there it is, plain and simple. There are also many things that can only be done with cryptic commands, the scenarios that you have mentioned are actually a piece of cake. Anyways, that was a very nice video and it is always nice to see your hard work being appreciated :) Truth is that I was heavily inspired by early Wendover videos and decided that the commercial side of airline operations is equally important and interesting to the flight operations one. So thank you Sam!
My buddy is a flight attendant with another major, so take this with a little grain of salt I guess, but they said a major component of the Southwest meltdown was that when flight crew call in to get new assignments they call into the same number that everyone else does… which, when you cancel 10,000 flights can be a little hard to get through on!
Problems will also arise for old software programs that have been written in old programming languages that are no longer part of modern IT education. Whenever the existing generation of programmers reaches pension age, there is nobody to replace them. This becomes a problem whenever the old systems need to be adjusted to operational changes.
So at Delta we use that software but internally we call it DLTerm. So let's say you want to look up a flight from say Atlanta to MSP (Minneapolis Airport) you would write on the system SATLMSP then it will spit out in bunch writing the flight time, departure time and different classes available in letters it's pretty archaic system but it's one even us devs don't want to touch because if you fix one thing something else breaks and most of the people who originally built these systems retired back in 2020. Fun fact it's quite common to leave something broken because if you fix that something else might break or their might be a cascading effect so if the thing broken isn't important or critical we just let it be.
You would not belive how much code that society is depending on is legacy code from 80-s or myb 70-s .. ( similar situation is inside all banks and credit cards ) .. And yes I have huge respect to software developers from that period ..
@@danicic87 Make that 1960s for both banking and the core of the GDS. COBOL for the win, all credit to the amazing developers and the IT staff who chewing gum and rubber band to keep it working
@@ariandeelsii sorry, you're giving too much credit to these developers and IT. I know quite a few "old timers". IT was very simple back then: call support, which was 99% of the time IBM. 20 years ago I had a teacher who complained that "now you have to do all yourself, in my day IBM was great, they even came without us even calling them, because the mainframe had an error and called IBM by itself". as for developers, we're talking about an extremely conservative industry with rules set in stone for decades. those systems are extremely simple (and because they're simple, they're so reliable). New developers have to work with the latest-and-greatest always pushing what's possible because that's what the market is about. Have to be inventing solutions with unclear and ever changing requirements. a coworker of mine, her dad worked at the bank in the 70s. he did the software in COBOL for the bank. the guy was a construction worker who grabbed a COBOL book one day and showed up at the bank when they were hiring.
@@amcluesent That is kind of the point nowadays though... Low sophistication means low susceptibility to cyber warfare, which is what you need with ICBMs. As long as they guarantee a counter launch before the enemy nukes hit the silos, that is good enough.
Fantastic simplification. After 17years in Travel Tech can’t tell you how many GDS integrations were cancelled due to ridiculous GDS developer licenses, etc. Amadeus, Sabre, et al made it too expensive to integrate so companies just do direct integrations with key sales channels. As you say, the APIs are old and severely limited, when combined with extortionate integration fees, it quickly becomes not worth integrating…
Even those pictures you showed of the British 90’s travel agent IT kit was still just a front for another ancient it system that was running on even more ancient kit in a dusty data centre somewhere near slough
Ah, but what happens when "good enough" is no longer true? I think that's where we are now. It is PAINFUL to travel now. I've traveled for 40 years now and the past four years I have regularly canceled flights, which never happened before. The terrible TSA lines, delayed flights, charging extra for a cup of water and carry-on luggage, the overbooked flights, crew not available, crew is late, paying extra to SIT NEXT TO YOUR CHILD, gate agents are the worst people to exist on planet Earth, the 90 minutes on hold with support, and terrible online booking experiences is too much. It's too much. I don't know if it's possible to incentivize better behavior, but I'm in favor of treating this like a crisis where the government helps get everyone on the same track. I mean, we literally had to write a law on how long airlines could detain us on the tarmac because airlines didn't care.
Air travel is practically dehumanizing at this point. I live 900 miles from my mother and we’d rather do two days of driving to visit her than a two hour flight.
@@abbysc417 This is why the US needs a high-speed rail network, it would free up the airports of all the small local planes that need to connect town that already have a rail connection.
@@HalbeargameZ Shinkansen could be an option. Siemens' ICE trains have also proven reliable in many countries (Germany, Austria, Netherlands, Russia, China, to name a few). Also, there is the Alstom TGV, which is comfortable and good for high capacity with its double deck options.
One of the problems is that the current market actively rewards the airline that can get you from point a to b with the lowest initial ticket price. Stuff that was included in the ticket like seat reservation, luggage, carry-on etc, is now an upcharge because it looks more attractive on booking sites to have a low initial price rather than a fare where everything is bundled in. One of the reasons why everything else gets cut or worse in quality.
I appreciate that Sam marked Midway Airport in Chicago instead of O'Hare, as Southwest only flies out of Midway. This attention to detail is what makes Wendover so great and reputable.
Actually, Southwest *does* fly out of O'Hare now. They added it back in February 2021. However, it still flies a lot fewer flights to and from O'Hare than it does to and from Midway.
Another thing that contributed to the SWA meltdown was the ramp issue in DEN. A manager sent an email to all the rampers requiring everyone work overtime over Christmas. The message ended by saying if you didn’t show up you were saying you quit. I heard 90% of the DEN rampers called off sick. While not the full cause of the meltdown it made things much, much worse for no reason. A lot of SWA employees complain that the days of Herb are long gone and that the company no longer treats its employees like people, more like commodities.
That brings into play the other issue that airlines currently face, being understaffed. If they had more employees, they would be able to put more staff on standby for when issues arise. With current staffing levels, a blip in one place can cause the whole system to breakdown.
This has to be one of the most amazing insider videos I've ever watched. I work for one of the largest retailers in America and we run on Mainframes. As this story explains, the Mainframe didn't crash and that wasn't the issue. The issue, or at least one of the main issues, was that the scheduling system only scheduled planes, but not the crews. And the rule on crews timing out wasn't accounted for anywhere, which made things even worse.
Modern mainframes are remarkable pieces of technology. It's a shame they get written off as outdated or legacy or the source of the problem most of the time.
@@dankdank1891 Mainframes are centralized servers, usually built by IBM nowadays (I believe Hitachi also makes modern mainframes but I know little about it). Mainframes are much more powerful than decentralized servers (the servers people usually think of when they say "server") and usually have greater reliability than decentralized servers. Some mainframes configurations have been running for decades without downtime, thanks to redundant systems and sysplex. Linus Tech Tips (surprisingly) has a pretty good video about mainframe hardware
@@dankdank1891 A mainframes sole purpose, including having a fully customized z/OS or a very stripped down version of Linux is to process transactions and store the data as quickly as possible. It is all about efficiency. While the hardware has gotten a little better, the software hasn't changed much, why rewrite code that was already optimized to almost the fullest extent decades ago. Not every company needs a mainframe, only those that deal with large volume of concurrent transactions.
Rental car companies are like this as well, or at least they were 10 years ago when I worked at Enterprise. The in-city locations had Windows 98 level GUIs, but the airport locations ran on command line programs. And it was a constant rotation of employees, so there was usually only 1 or 2 people who had that senior-level experience because they were permanently based at the airport.
“Command line” is missing the point - CLIs can actually be very efficient (as also the video remarks). The main problem is the backends behind them. ...And the lack of education so younger employees can't use the proper interface and need to use the graphical layer which can't actually make use of all the functionality.
I work for a legacy brand airline and it’s amazing how old the technology is, just like you showed. But also like you show those of us who have worked for a few years know how to get ourselves around the system, the way regular passengers can’t.
I'm working for an airliner as a developer and we use Amadeus. We are moving to NDC and while better it causes a lot of headache and takes ages. Don't get fooled by the "new" in NDC because it is based on XML and feels already outdated before it's implemented fully. Really nice video, very well explained! Btw, 16:11 takes away a common misconception that we are dynamically changing the prices or based on individual customers. There are only a certain amount of slots for a price so if they are all sold the price increase, not because you are looking for a second time to the same flight.
What's this new-fangled JSON stuff, right? Now try explaining to Google they shouldn't pin TLS certificates ... Everybody doing NDC is doing at least something wrong!
@@petergamache5368 thing is they are extremly slow to build ne standards like NDC because so many Airlines have a say in it that it overcomplicates the whole process. And because of that, NDC is the next best thing.
Worked for KLM customer service, the onboarding alone was 2-3 months just because the booking system with those codes was so outdated and difficult it took a lot of practice. I still have nightmares of codes like AD15OCTAMSCDG and shit like that. We had to do that manually while we had customers on the phone to order tickets as well. Boy was that a shit show. From my group of 15 new hires 11 quit (incl me) after the 3 months onboarding period, and the outdated and annoying system was a contributing factor for all of us and one of the root issues for lack of motivation, high pressure and even stress for some.
I would much rather enter AD15OCTAMSCDG and have instant response with accurate availability than fill in endless boxes on a slow fiddly online form and wait for pages to load and see half the picture .... Just speaking as a Amadeus native user dealing with corporate travellers... I guarantee that making bookings, changing bookings, re-issues etc take a fraction of the time compared to NDC where , when forced to use , we grind to a halt. But yes, I've had my laughs with KLM call centres ... "married segments" HAHAHA
Only lasted 8 months on that posting. Typing out entries can be confusing at times but when you got to talk to a passenger on the line it gets more confusing. _I'd rather work on the ramp servicing aircraft than go back to Airline call centers again._
As someone who, on the daily, works with Amadeus reservations and Altea CM, it baffled me how easy things became after a while. At first it seemed like everything had its own entry and to some extent it’s true, everything ticket agents do has a backbone to the basics of creating a PNR.
It's not difficult, you're just not suited to it. Hundreds of thousands of travel professionals use it every day. It's fast, it's accurate and it's extremely powerful. Airlines have been trying to build GUIs and alternative systems to replace Native for over two decades now, and nothing comes even close to replacing it.
A good presentation, and ultimately the key phrase is 'what's good for the stockholder ...'. Every corporation will end up cutting it's own throat because it can't see past the quarterly earnings that are 'good for the stockholder'. It truly amazes me that their stockholders, while earning a little less, will earn far more in the long run if they take the long view where public good and futures investment is considered.
it's about the ceo execs and boardmembers not stockholders. If a ceo comes onboard a profitable well run company..cuts salaries and benefits, slashes the maintenance and puts its departments on starvation budgets profit jumps short term as the company feeds off itself. 5 years later his reighn has been very profittable, he the board and the execs have recieved massive bonuses..he leaves and the next ceo takes charge of a broken down barely functional organism in it's death throws and the profit and stock prices start to reflect that
This reminds me what happened with internet in early 2000s. Initially we built many interoperable protocols for emails, chats, file transfer, voice calls that were software agnostic - you could be using one software program on one end and a different one on the other end and they worked well together. It worked remarkably well in 90s. Today they all except email died out and been replaced with non-inteoperable corporate cloud software as service products - facebook, google drive, onedrive, etc. It was interesting to learn that in airlines industry this olden approach survived for so long. In nutshell the problem with interoperable protocols is that once they get adopted it becomes too hard to change anything about them to the point where they get replaced by private closed propriatery product that is just many miles better.
I feel like the only semi interoperable software I have is blender for the single reason it has a bunch of file formats it can save as or load except for whatever the he’ll hbaked is
If you know about raising money, #1 money is infinite #2 only if the problem is big and obvious enough is it worth spending #3 when you do spend, make sure it's super good and novel so it lasts a long time. So it makes sense these are the patterns in airlines.
I was a consultant for an airline two decades ago. Their IT systems were 30 years old, allowing only four simultaneous connections for some important things. Although it must be said that the bunker for the old mainframes was big and awesome.
I used to train agents on Worldspan 20 years ago. Worldspan was basically TWA and Pan Am's PARS system and replaced Delta's Datas II. It is exactly the same now as it was 20 years ago. So a system written for TWA and Pan Am is still in used along with SABRE (which is even older). That should instill confidence
These systems were developed in the 60’s so they’re now 60+ years old. Unfortunately the ‘new’ NDC (XML) interfaces need to be backwards compatible so while interfaces may have changed, the underlying logic/limitations have not.
This is one of the single MOST IMPORTANT VIDEOS for travelers to watch. Helping them (and myself) understand why booking anything but a highly-typical itinerary can be a headache and seemingly extremely expensive. Thank you Wendover for making this video 🎉🎉🎉
The most widely ignore software design rule is to fail *early* - better a small, localized crash that can be fixed by the developers actually responsible and while they remember all the context, than keeping it limping along with a bunch of squelched warnings and eventually cause a chain reaction of other failures.
Honestly? It's best to just focus on fault recreation. Never ever try to 'handle' anything that isn't an exception generated by an unreliable system (connection timeouts, retry, etc) Number one priority, record what happened (public entry point) with which parameters, the stack trace and every version number required. Ensure your build process stamps unique version numbers write a new unit test to catch the fault before deployment, patch the code and turn the handle on the release process. Don't have one of those? God help your damned soul, because I wouldn't touch your company with an infinity barge pole.
Very good summation of the Status Quo. I liked your definition of the current airlines situation as "Reverse tragedy of the Commons". There are two solutions: one is the mandated government minimum standard of operation; two is setting a voluntary inter-airlines independent commission who sets a standard that every airline is obliged to implement. Servers and applications have come down in price since the IBM System/360 was running that command-line interface (don't ask how I know). Thank you for the video. Pretty much every aviation channel on UA-cam has made an oven-ready video on Southwest meltdown, but your Wendover video is the only one which had detailed the mechanism of the failure and the relations between the various software programs (apps?). I agree, the communication protocol may deserve the ½ century upgrade. Greetings from the UK, Anthony
So, in other words, SouthWest suffers from what all big businesses suffer from: PROFIT FOR THE SHAREHOLDER over all else at all costs. Profit for the shareholder over safety. Profit for the shareholder over employee pay. Profit for the shareholder over customer comfort. Profit for the shareholder over all else, always.
How convenient for timing. I went on delta airlines yesterday and mid flight, their entertainment systems(which run on linux) had to reboot because of malfunctions
Fun fact, Linux is the de facto OS for Supercomputers and Financial trading servers so the OS is not the limiting factor for why Delta's entertainment system had to reboot on your flight. Delta could easily make their entertainment system more robust so that it could handle software failures and hardware malfunctions while providing 100% uptime availability. But they will never do that as it's an order of magnitude more expensive to do so and they have no financial or PR incentives to spend that kind of money.
This problem of not adapting to newer software technology solutions is not just limited to the Airlines. You could also find it in a other legacy businesses. Take for example, Neiman Marcus, an old-school retailer. When I used to work for them right before the Pandemic, they were still using command line software developed in the 90s for their employees to clock in and for their inventory and logistics. Even though the computers were Relatively modern looking desktops. For example, we would open up one of the proprietary programs clicking the icon on a windows desktop, and the program’s window would pop up displaying a black background with green text like it was 1983.
Pretty good video. GDS are still using cryptic entries, but Amadeus is not any longer cryptic native since it migrated out of TPF, it is just backward compatible as travel agents are actually way faster using cryptic entries. And yes GDS only provides interlining routes as possible routes, this is however a commercial limitation rather than a technical limitation. Basically your itinerary is under the same e-ticket, and whoever is the airline in the e-ticket is the one in charge of bringing you to destination. But yes passenger could take a combination of separate tickets for the same itinerary. And several companies saw this as an opportunity to create "virtual interlinings". Comes with some drawbacks though: you have to pick-up you luggage and re-do the check-in mid way, and you are on your own if the connection is missed. "Virtual interlinings" companies try to mitigate those drawbacks with agreement with the airports for your luggage, and with insurance in case of missed connections.
The #1 asset of an airline is their safety record, the #2 is their reliability record. You would have to be crazy to buy tickets on SWA the week of Christmas 🎄
No, it's proximity to where people live and low costs. Ryanair is a universally hated - not disliked, no, outright hated - airline despite its 100% safety record, in over 25 years of operation they never had a single lethal accident. KLM, which was involved in the Tenerife disaster, the most horriffic air crash in history, is a lot better liked. The reason people fly Ryanair instead of KLM is that you can fly from easy to reach local airports. Within an hour of travel by car, 3 Ryanair bases are located from where I live: Eindhoven, Düsseldorf, and Weeze. The closest international operations base of KLM is Schiphol, which is twice as far away. My experiences with KLM are rare but were very positive. I still fly with these cunts from Ryanair because Eindhoven is just too close and too easy to travel to (even by train and bus it's just an hour away from here) to justify travelling twice as far and paying a slightly (or significantly) higher fare just to avoid them.
This was by far one of the best, most informative presentations you’ve made to date. And I watch every single one. I was seriously schooled. And I worked for United Express doing baggage service from 1998 to 2000.
Great video, I was under no illusion that the IT systems in the airline industry were old and complex, but it was neat to learn specifics. Love this sort of "behind the scenes look"
I was sitting in the airport during the FAA system restart. It was really weird to get notifications that our flight was delayed right before the news broadcast behind us said all flights were grounded. No one knew what to do or whether their connections would be there when they landed. Even the flight crews were like "we're going to board and try to take off at 9:30, but no promises we'll get off the ground"
The US needs to privatize its air traffic control service. After Canada privatized its service, it became of the best if not the best air traffic control service in the world.
Ancient IT just saved Southwest from Crowdstrike. It's hilarious but points out the value of heterogenous IT systems because when everything runs the same OS then some vendor pushes an update a fecalstorm ensues.
Retired from a large software company 2 years ago after 32 years. When I was first hired in 1990, there was then-antiquated airline software that was associated with a product I worked with. It was a real chore to find skills to mend anything that went awry, or to scaffold a system in order to test fixes to the archaic code. I believe that the software is still used at many airlines, and it looks about as friendly as the 3270 green screens in this vid. Memory was really expensive in the days when the software was created, hence the brevity of presented data. Amazing that such old code works as well as it does.
Another great explainer video! As a veteran of the travel industry with over 30 years experience, I'm glad you packaged this conundrum in such an easily consumable way. And as a technology partner to airlines, we really hope to break down some of these inhibiting stalemate factors. Nice one, Wendover!
CLI are quicker a lot of the time than GUIs, at least for trained employees. I did 2 years as a shop assistant and we used a CLI, boss pressed 8 buttons and could pull up any order. It was insane how shortcut keys worked.
A) you could argue that the much increased training and adoption time may outweigh the benefits of a CLI, if the GUI is actually well designed, and decently speedy B) the two should not be exclusive. I still hate how strategy games figured out decades ago that you can have an easy to use GUI and a million shortcuts at the same time, and still make it work (with, let's be honest, minimal dev time overhead), and a big chunk of commercial software still doesn't let me Tab through options, or hit Enter for OK on a pop-up.
A bunch of the ick here is using a CLI built decades ago for trained humans, as the machine-focused API. "Easy for humans to read" rarely translates into "easy for machines to read", leading to brittle code that's shoved in a box and nobody dares touch it ever again. A better design is that the CLI is simply one more interface over the same machine-friendly API, which can be designed to accomodate new concepts without risking breaking every existing client. CLI versus GUI is a much falser dihoctomy than is usually realised, and I suspect frequently turned into a console war of character cell terminals from the fucking 1970s, versus actually admitting you're using a bitmapped display to emulate that terminal. Modern twists are possible, with results displayed graphically (eg, numerous RDBMS frontends), and maybe even using graphical rendering to call out things like the boundaries of values rather than text. And even then, the problem isn't terminal vs bitmap. It's lazily developed--out of laziness, cheapness, or being in a hurry--software that uses bottom-tier design, without having things like commands as a concept in themselves, or reflection that can be used to automatically derive a scripting interface, or even being able to visualise data in a fairly raw way.
Telecom had the same problem: backbone of most of the systems was written in Cobol, & you could typically count the number of people on staff that knew Cobol on one hand. When those people retired & something's not working correctly, the hunt was on to find one of the remaining people...to save money, telecom management would "kick the can down the road" rather than fund modernization (too costly), with predictable results.
There's a few other places where this "if only you knew how bad things are" situation is an everyday reality, banks for example with legacy cobol & fortran systems with more modern solution on top or to the side of the older stuff. Utility like electricity & water in some countries are like this too.
I can't tell you how long I've wanted to know this and how much I enjoyed finally learning about it!! That being said, this shit is so complicated I could barely keep up with it and I have no clue how you were able to grasp it enough to make the video itself. Back in the early 2000's I called the airport to look into a flight that someone had booked for me, I was lucky enough to talk to someone who genuinely cared enough to want to share some useful info with me. Again it's irrelevant now, but at the time it was priceless. She gave me a number to call if I wanted to book last minute flight via any available cancellations. That was basically like being given the most valuable thing ever in the realm of booking flights. It's a shame that I was never able to make use of it, but it meant something to me at the time.
Two things of note 7min in. We don't call reposition flights "passengerless deadheads". Deadheading is a term we'll use for crew about exclusively. We'll call them reposition ferry flights in the industry. Also, the maps image of the FAA headquarters with the ground stop that was issued wouldn't have happened at that location. That would have been at the Air Traffic Control System Command Center (ATCSCC) in Warrenton, VA. It's right next to Patomac TRACON.
This was one of the best explanations of legacy technology and the problems it causes. I've worked in enterprise for my entire career and it was initially shocking (and still terrifying) to see how much critical technology is stood up on tooling created so long ago. They're all hard problems to solve for sure, but this kind of problem is going to happen more and more across industries.
as an employee for an airline mentioned in this video - they have tried to replace these command line tools with fancy new front ends, but we always go back to using the good old green and black screens as they just work so quickly! Long live command line interfaces!
Yes, finally, I work at a travel agency that's specializes in the marine sector, I honestly from the bottom of my heart to the depths of my soul HATE THE FRONTENDS give me my cryptic line and I will book you in 30 seconds to wherever you want, sure, for the newcomers the cryptic entries are a pain to learn BUT once you master them you cannot in good conscience use the frontends. Let's say that I use sprk to book from ATH to FRA, it LITERALLY takes me 2 minutes to create the reservation, on Sabre I need no more than 30 to 40 seconds plus I have all my quick keys programmed... NDC is just a fancy way of jamming more ancillary services to the pnrs it does nothing to help with the reservations... Actually it's even slower since for each reservation the GDS must download all the information that it's not needed...
@@nicknellas1287 One of the main problems with modern software development is that non-technical people (who are often in charge of the software development) have trouble understanding anything that's not a graphical UI. I imagine that this explains a lot of the useless new frontends: they are easy to sell because the CEO can touch and feel them, even though they don't fix any underlying problem with the ancient backend systems.
Worked retail in 2000. Had green screen look up for inventory, price, wholesale price, reorder etc, way more than a sales clerk would need. I only knew a couple commands, but it was damn useful and fast. If it said it had it, we had it. Then they replaced it with heavily locked down Windows running on comically underpowered hardware, as in typing would lag behind significantly. Everything took forever and told you nothing. Insult to injury the colour bit depth was set to 8 causing eye bleeding dithering that, if switched to 24bit manually via some workarounds, ran at least 10 times faster making it useable and looked as it should.
I knew as this company started cancelling flights that it had to be a lack of upkeep for tech, damn shame for all of those missed flights but I sure hope they fix it soon.
I promise that the last 15 years the IT dept has done nothing except warn higher ups of an impending failure of the system, and the higher ups likely said 'We'll think about it' while stretching the backbone of their company on box fans pointed at server racks in rooms where the AC failed 5 years ago.
Always prepare for the worst, because the worst WILL happen given enough time, and however much it costs to prepare, I promise it'll be cheaper than waiting for the worst to happen.
Yeah, at this point I'm almost convinced politicians won't do anything until the system actually collapses. So whenever that happens it'll be a year or two of pain for whoever needs to fly somewhere, as the industry scrambles to rebuild their entire infrastructure from scratch. 😛
Technology is never the problem. Incompetence OR Short-Term optimization is. In this case: Both
at this point front line customer facing departments have joined IT's chorus
The trouble is the IT department generates no income but has big capitals and overheads. Shave some of that and hello promotion! I've been running the IT department at my small company for the past 11 years and the only message I ever get from the CEO is COST COST COST!!!!!
@@chuckygobyebye Interesting, can’t you communicate with the ceo that IT has its benefits and investment returns?
Airline consultant here: this is INSANE. The fact that you included information that even I didn't know is beyond amazing. I only today realized how much research goes into these videos like...insane, even the internal tools wow. Great great job!!!
It really is amazing, even if you figure in that there is a team behind it.
don't forget the passengers exploit ATPCO through usage of VPNs and get a fare for just 310 dollar for a long haul flight from let's say Dallas Fort Worth DFW to Amsterdam, without VPN and booking from the netherlands it's 1200euro's but connect to a VPN in Mexico it's just 400 and with a discount voucher you get insanely low prices, and... say amsterdam can erupt into chaos it get cancelled but not necessarily due to weather but due to a technical fault caused by icing during departure from amsterdam because they didn't de-ice a really frozen Flap pod. then it classifies as a technical problem. flaps problems and one side of te flaps working good the other not can result in a deadly crash. because of 5cm of snow but freezing -20 overnight in march..... so you get to fly on an airfrance flight on an A350 to paris with a free TGV ticket from CDG to Amsterdam, the fuel economy being way better then their A330-200's of KLM you get to that 310mark and choose your seeting for 10euro's extra. and the passenger who has to wait for the airfrance flight to depart 3 hours later, but upon arriving in CDG just hop on a TGV and transfer to a thalys at lille'europe or in brussels On a amsterdam express which will be operated by highspeed trains as well (the ICNG), the eurostar coming FROM london doesn't require lengthy passport checks. a thallys must be reserved and i dont know if they run from CDG directly to amsterdam let's just say a TGV + ICNG is cheapest for the airline. and why not brussels? well brussels airport is more inland and a flight to there would deviate much further actually causing more problem and delays
Part of it is a true story... i experienced. especially the crazy price differences and refreshing pages cache's and virtual machines with VPN's to book tickets, amazing. and yeah rerouted 500south f your destination with a highspeed train a thalys from gare du nord would take 2hoursand 40 minutes with more legroom and a awesome 300km/h experience, the only slowdown occurs between brusselsand antwerp were they they do 160km/h but can do 200 with ERMTS enables but not all trains i belgium and the Netherlands have ECTS that is used with incab signalling. from antwerp to amsterdam it's a 300km/h ride again with a stop in rotterdam. and if you live closer to rotterdam than amsterdamm or even closer to antwerp like TErneuzen or near or in breda in antwerp or brussels you take the brussels express and in HSL zuid in the netherlands thers a point were trains can exit high speedline and get to breda very quickly..... it's just amazing parnetership with rail AirfranceKLM SNCF SNCB DB and NS that allows that option as 5CM of snow is not gonna stop a train
Not paying attention would do that, yes.
100% agreed. I worked in consulting with Sabre and Amadeus 20 years ago. The detail in this video is extremely impressive.
@Teamgeist no way to verify without spending just as much effort... and no effective reason to doubt, in my case.
The channel seems pretty legit, this does not strongly affect me, so I can just choose to trust here, without much concern.
The fact that literally 40 years after taking my first flight I still can't purchase an airline ticket using my full and correct legal last name because it contains a hyphen is mind-blowing to me. That really says everything you need to know about the airline industry's IT systems.
Wait what? You can't get a ticket cause ur names double barreled??
@@Evangelionism Can confirm. Can never figure out whether they want me to leave out the hyphen altogether, replace it with a space... and oddly, this isn't on most airlines' FAQ pages online, despite hyphenated last names not being THAT uncommon.
@@inspiratron That s interesting and pretty ridiculous man. My first name is double barreled (although I usually spell it in a way where I don't include the hyphen, sometimes even for legal purposes) and my family names are double barreled. I haven't had any issue traveling before but buying things online using my bank card can be a bxtch sometimes! Couldn't buy myself Discord nitro using my card cause the hyphenation of my family name made it hard to tell whether to use a space or a dash (neither worked).
Oooooooooh yeah, don't get me started with it lol
Alex accented names too.
Like Ólafur, won't work. Either it'll reject, come out as Ólafur, or lafur.
As an IT professional, corps VERY rarely replace or upgrade hardware until it literally is about to burst or already burst. Its VERY frustrating. Ive been used as a scapegoat for this shit before too.
The funny thing is individuals basically do the same. The most up to date tech I have is my Gaming Rig, all so I can play skyrim with mods.
And trying to explain that the cost of an eventual meltdown is going to be way more painful that an organized, preplanned upgrade just goes in one ear, out the other...
@@eldermoose7938 now that you mention it, my graphics card is a 1070
But hey it plays Hogwarts legacy just fine so it can't be too outdated
Me too dude. The Bean counters simply don't understand that the cost to replace obsolete ticking time bombs in the environment is nothing compared to the monstrosity of a headache and of a cost that it will be to try to get a company back up and running if a piece of the critical infrastructure goes kaput. Now you're paying all kinds of emergency fees from consultants and emergency rush jobs from other places to get your hardware faster or whatever the case might be, all while the brass are screaming at you to get it back up and running
@F L
Quarter by quarter, short term gains. It’s not sustainable in the long term. That’s the biggest flaw of capitalism.
I no longer consider Southwest to be a "low-cost airline." Its fares are often right in line and sometimes higher than the traditional big carriers.
no longer low-cost
still low-comfort?
I think it depends on the route. Where I live they are still always 50% of United. But other routes are the same
Agreed, I never understood why they were considered a "low cost carrier." They're never the cheapest, and they never go where I need to get to.
"Low-cost carrier" is a business model description (like network carriers, charter airlines, and others), not necessarily a price observation. Lower turn-around times, uniformity in aircraft types and seating classes, direct instead of hub-and-spoke flights, etc. are all intended to reduce costs, but don't guarantee drastically cheaper prices than other airlines (especially since these days, many non-LCC's are still influenced by some of the LCC philosophy)
I'd say they're Low Cost but not ULTRA Low Cost like say Allegiant/Ryanair etc.
As someone who works in IT, it is completely unsurprising that this is the case. You'd be appalled by just how many vitally important things run on highly outdated tech. This is usually because the cost, manpower, and time required to replace it is often absolutely monumental. So company bigwigs stick with "well if it ain't broke, don't fix it." Even the pentagon still has critical systems that run on windows 95 and 98.
Not saying it's right or the best way of doing things, but there's usually more to the story than just pure money or greed in these cases. Nothing in this field is ever as simple as just pressing a few buttons on a keyboard or unplugging a device and plugging a new one in.
Yes it's not easy to implement new software.
Yeah... We work on systems running XP but when I worked on IT airline systems that was a whole new level of ancient. OS2 , token ring, and whole airlines running 56k links
Yeah it’s incredibly naive when people say “just upgrade the system!” As if it’s so simple and cheap and can be done in a few hours or something.
Exactly. I work in IT for an airline and your comment is spot on
Sometimes updating to the newest software can cause more issues short term than the old software was. And in areas like engineering, software like cad used to be a pay once and own it forever. Now its exclusively a subscription that costs as much as the original own it forever price so no one swaps out the old software going as far as to setup virtual machines to run it.
I work for a software consulting company. Basically my job is to write custom software for businesses. So naturally, I deal with technical debt on a daily basis. It can be very frustrating at times because we are not in control of whether or not the hulking old system is replaced or improved. That's in the hands of our clients. We can always advise, but never command.
However, when we _are_ actually contracted to replace or improve old systems, my job suddenly becomes amazing. The level of impact I can have is staggering. My work can affect people all over my state, or even all over my country. I love doing this kind of work, but I have to be asked to do it, and given the money and time to do it right.
If you're out there and you're in charge, take a good, hard look at your systems and ask whether they could be considered modern and efficient. Ask what the cost of failure is. Then do something about it, before the whole thing caves in.
I had a new client 6 years ago that had been plagued with constant minor failures and quarterly major failures. After evaluating the entirety of their infrastructure, procedures, and software I recommend a complete overhaul. Everything including the wiring was replaced. New software was deployed, and new procedures were written. It’s been 5 years since everything was completed, I only see them annually now to check in. No issues are ever reported. When you do it right and spend both the time and the money, things just work.
@@Veritas-invenitur great to hear it's doing so well. It gives me hope.
Borrowing at the Bank of Bodge and then paying back with interest.
Are you sure we need to upgrade? This has worked fine for years. What's the ROI? How do we know we really NEED to upgrade? Ha!
This is a wonderful anecdote, and I also think an example of how important Systems Design is of the most critical things when creating long-term solutions. I also have a similar responsibility with a small data team and think about how doing things with effort can have a positive impact not only on the company but also thousands of customers is rewarding and motivating.
As an ex-travel industry employee (thanks, Covid), seeing that GDS screen gave me some PTSD. Travel agents call it 'flightmares' - stress thinking about doing something wrong in the GDS system (which can result in expensive fines from the airlines if you break booking rules accidentally that the system allows) or having to do some major re-booking. I don't miss flight bookings at all!
As some of the of the Airlines work to upgrade their outward facing system the GDS starts to look more like a bunch of json data. I guess it explains why I've heard it called the "database."
TA here and i completely agree its a complete mess , also having worked on the airline side. i understand both and both are a mess
This is crazy to think about. The only reason I re-entered aviation is because of covid. So many commercial airliners need to be stored because flights are still down, which is the reason I even have a job right now (I'm a mechanic at a long-term storage facility)
i heard of the problem of "screen fatigue" from military enthusiast circles
its mostly about how long a radar operator could stare into a screen until his eyes feel tired and begin to affect his performance, which is crucial in detecting enemy planes, missiles and stuffs
seems like its not really unique to military now
Seems like a perfect work for AI. (not joking, Ive worked some administrative and managment jobs that no one wants to do and everyone would probably highfive if there was an AI to take over, letting us focus on more "humane" tasks)
Thanks for your proper voiceover with natural pauses. So sick of lazy UA-camrs who splice a thousand edits together, chopping out all pauses and creating a never-ending barrage of words where you don't have a chance to think or absorb the info.
This channel has a long list of great vids like this. Quality narration, info, editing.
I finally retired from IT after 40+ years in it. We would beg and plead with management to spend more money on staffing, upgrades, etc., all to no avail. Then, when something finally did break, it was, somehow, all our fault.
Of course, that never stopped them from spending money to implement the latest fad in software development or whatever and trying to shove it down our throats, thus making it even more difficult to do our jobs.
Feh!
I'm in IT and a pilot. The cost of upgrading the systems would be horrendous. The amount of FAA and federal red tape and testing would be insane. It would literally have to be all the major Alliances getting together and crowd funding a standard system its that bad. Also, keep in mind Airlines aren't about new technology, they are about stable, and supportable.
@@ooo_Kim_Chi_ooo I don't think you get the essence of comment above - it's not about new technologies or upgrading everything at once - it's about continuos improvement, which is impossible when managers have mindset of "do not touch if it works". Simple as that. Yes, I can agree that system so safety-critical on which many other system depend has a lot of inertia, but that's only argument against revolution, not against evolution. And last, but not least: the amount of money developers are begging for is magnitude lower than money spend on investments in other branches of this industry.
@@proosee Its just not possible though in the airline space. I mean people easily forget that in the 80's planes were literally falling out of the sky and when Airbus introduced all their wizardry it didn't go off without a hitch. The airline industry has been bit hard by new technology whether it be computers or hardware. They are extremely slow to move on anything and approach more of an it's working perfectly now don't touch it until we need to approach. When you are dealing with lives and money at such a scale their willingness to be up to date is lessened by the fact that they really have no really reason too other than its old and shitty to use.
@@ooo_Kim_Chi_ooo still: it's not "willingness to be up to date" - literally entire video is about weak spots of this system. I beginning to doubt you ever integrated with legacy system, because if you say "shitty to use" is the most significant problem then you are doing those kind of integrations wrong - you can easily create some intermediate layer to isolate from it and never thing about it again. The case here is that this system is unreliable, slow and faulty - and yes, there are new technologies that solve those problems, but word "new" is not the most important here.
@Guy Montag That's the trick, all right.
I worked for about 5 years with a travel agency, and got really familiar with Sabre GDS. The tool itself was quite marvelous as a new agent, but as the years went on I started to notice the huge amount of limitations it had. For example, the entirety of a corporate account for one of the biggest oil companies in the world could be messed up quite easily by one single agent unticking a box while looking up a PNR. Happened many times, and I swear there was never a change on the GDS to avoid delaying thousands of passengers from booking their flights.
Well as long as that oil company was getting fucked I have no problem with that.
Hmm, what box is that? Granted, I don't use the graphical interface (I've been on this GDS for 30 years, don't need it). Are you referring to a command that allows the editing of the 1st level star?
isn’t that the opposite of a limitation? proves that it grants too much rather than too little,
@@gfjfjufidi2880 no its a limitation because of how few people are trusted with systems that don't have sanity checks in place.
I'm wondering how much the giant oil company got discount % from normal fare? I believe they are mostly travel in J
When I went to college in the late eighties, one of our profs had experience in the airline industry and helped with Sabre. Fast forward to 30 years later, we are still using the software he worked on. How sad is that.
Wow
Damn. 🤦
that would be demoralising. i could never work at a place like that.
US Airways had updated their reservation system and the agents liked it before America West bought them. America West brought Sabre back, and then they stuck with Sabre through the merger with American...yay!
I'm a tech for a company that provides IT support to airports all over the world. Sabre, IMAGE, and whatever the heck Delta Airlines uses are incredibly dated and clunky. They look like an app running in Windows 3.1
My mom worked for Air Canada in reservations and bookings in the late 60's and early 70's. They had one of the most up to date systems at the time. It sounds like not much has changed since then, and that she could literally do the same job 50 years later with almost no additional training. (She still remembers most of the 3 letter airport codes too!)
What has changed is their profitability.
3 letter airport codes never go away. they are burned in my brain and finger muscle memory. Its hard to type or and not ord.
4 letter codes are better.
@@wta1518 why do you say so? Personally I like 3 letter codes just because I'm so used to it, and they usually have some resemblance to the location of the airport.
@@SKAOG21 So do the 4 letter codes, at least in the US they're typically just the 3 letter code with a K at the beginning (eg. SFO becomes KSFO).
My father worked in ticketing many years until retirement. He used to work with an Amadeus terminal and type the most complicated reservations on the command line by hand. He then was unable to work with a windows computer or use the mouse 😂 This brought good memories, thanks for the video.
Travelport Galileo > Amadeus 😉
GDS system babyyyyy😅
My mother is the same, she'd destroy me on a command line but can't figure out how to drag a file between folders lol
Ah yeah... The old Story of "technical dept". Technicians always warn and management never listens.
Always have stuff in writing, so if someone is looking to blame someone they can't blame you.
Valid comment, but it's "technical debt", not dept.
Curiously, I've seen a ton of people get it like dept
I am in technical management, and all I think about is technical debt. Today’s technical debt, is tomorrow’s headache.
It's not a simple case of management not listening. There's no incentive to listen as the video says. You upgrading mostly favours the competition then you. American Govt or international airline organizations could easily fix this by mandating the new solution within a year for all airlines. But they aren't doing it.
It's not just management, it's the accountants too
This isn't just an issue with Airlines, as someone who's worked in enterprise IT for Large companies and Governments for a few decades now, everything may look nice, but almost all system backends are running on either an ancient system, or ported over from such (The old software just moved to newer computing systems with a new-ish user interface, but the backend is still talking in what most would call "dead" protocols and database languages. Cobol, Fortran, etc.) Same goes for those "fancy" systems that healthcare use, it's all based on a 1960's protocol (M, or MUMPS)
The problem is they become so used, that nobody wants to be the person that changes it out in fear that it'll break things more, and or deal with the downtime involved.
On top of that most "new" systems are being developed by "Cloud" software companies, rather than software you can own/run, customers are being "locked" into a company that can... once under contract, be able to hold your data for ransom and charge whatever they want. (This is the main reason behind the "cloud", it's all subscription "locking" in customers, not about saving money or reducing support needs or cost).
i agree with all of this but especially your last point. between all the different saas/cloud companies that either go out of business, get bought out & slowly scrapped for parts by a larger company, or completely overhaul their product/company focus out of nowhere-- making the decision nowadays to try to upgrade legacy services is a real rock and a hard place.
and that's not even to mention the challenges around sourcing new hardware these days (ironically because you'd be competing against those same cloud companies massively scaling their data centres)
Can't companies create a Linux system or they should look at India. There are lot of firms who will make non cloud software that will give ownership to company
The difficulty with the legacy code in Cobol is the lack of tools we became used to in the last decade. Automatic deployment and thorough automatic end to end testing are a hell of a drug. And even more expensive when to be introduced after the fact to a system that is about six decades old and full of tightly interwoven ideosyncracies noone knows fully or may even be in a position to start to understand.
Re-implementation is going to be expensive, but for what gain?
@@sorryi6685 yeah, Linux is great for servers, but no company which isn't IT focused and wants to be successful will touch Linux.
With Linux, you as a company have to operate and maintain all of it, while you as a company do not have the expertise and not the resources to stay up to date. The same way a production company doesn't run their own haulage subsidiary.
I agree with the first part, but not with the cloud/saas.
Does a production company operate their own haulage subsidiary or have inhouse Architects? No? Why not?
Bc it's not your business model, so you should only keep the minimum inhouse infrastructure, the rest you try to outsource if possible.
The bigger problems are company mentalities, who think that the IT system has to fit all processes, instead of finding a middle ground - pick the software which fits your processes best, and adapt your real world processes as much as possible to the IT landscape.
Cloud means you outsource cyber security (only a few companies can afford an actual, decent system inhouse) as much as possible, you don't care about software updates which can stop your production, you minimize the hassle of finding good IT guys (especially if your company is in the country side and a bad IT guy is always worse then none), you need less expensive hardware to run demanding software, lower capital cost and better scaling (classical example, Photoshop - one time licence is ridiculously expensive, a monthly payment plan which flexible in-/decrease is create. Or how big is your server supposed to be? Doesn't matter, take the small one and book additional capacities when needed instead of overscaling your IT hardware landscape)
Has cloud downsides and do you need external consultants to support your decision (not making it for you) which one to pick? Obviously.
But these negatives dwarf by the alternative - keeping everything inhouse.
If tax dollars are used to bail them out, they should be held to a general standard regardless of what it will cost them. And that's the end of it.
If they needed to be bailed out then it should be a publicly held company that is sponsored by the government and run by the people, not a corporation, since the government essentially bought it.
@@t.j.vellinga6225 Unfortunately, many of these guys are "too big to fail." Boeing for example, they have been a shit company that ignores safety protocols and all about cutting corners to save cost, yet they keep getting gov contracts to keep themselves afloat. It's because Boeing is one of the largest employers in the US, and if they went belly up, tens of thousands of Americans will lose their jobs. Boeing is holding the American people hostage to force bailouts and contracts from the gov. It's the same thing with many of these big corporations. Their argument is always "If we crashed, we will take down tends of thousands of jobs with us and it will hurt the US economy more than just giving us a few billion dollars to get bailed out."
@@t.j.vellinga6225 Because the government is really good at running things (－‸ლ). Just let them fail.
In the US the airline industry is already subsidized by the government and also has contracts to fly US mail. Any more money given to them will go straight into the airline executives' pockets.
@@ironroad18 Exactly. Remove the execs and socialize the airline industry.
Google Flights doesn't do a GDS lookup. It buys availability data directly from OAG and consumes AVS messages from airlines directly and use a much smarter data structure internally to quickly evaluate farings.
This is my understanding as well, was surprised to hear otherwise in the video.
Yea, I found it weird that the way it was described as being bad is that "behind the pretty UI is a system that joins data from multiple APIs" like yea... that's the entire IT sector right there lol
Iirc Google bought ITA Software, which powered this kind of search for a while and even supported some small airlines, maybe just one Caribbean one. And they got a ton of flack about anti trust from a lot of air travel businesses for basically side stepping the GDS.
@@getellied I was thinking the same thing, but one of the key points was that those APIs haven't really changed in any meaningful way. I'm guessing GDS is mainframe based, which it's pretty hard to find developers for now, and architecturally change would be incredibly risky. Typical enterprise risk avoidance, where they avoid risk so hard that they end up in more risk in the long run.
That was also built by about 1/4 of the people who were at the MIT AI lab in its heyday. (Disclosure/Source: I worked for ITA, but not on QPX.)
As a private pilot and IT professional, I've been following this story closely. Thanks for breaking it down. What a lovely morning coffee watch.
I thought to myself: what’s a ☕️ ⏰ ? A watch you wear made of coffee beans? A coffee-themed watch? Oye. I haven’t had my coffee yet clearly!
as a person who watches youtube videos, I watched this video today.
I went to flight school through college and got partially through , then dropped out to work IT haha. Loving working as a field service tech right now.
The college program (Eastern Michigan University) was horrible; I was told something like 48/60 of students in my year ended up switching out of the flight science path and got av management degrees or MBAs. We had to take multiple semesters of aviation-oriented classes that were almost entirely useless (stuff like weather science, adjacent to aviation but not applicable in a pilot's career)
It's hilarious to me how important the aviation industry is to society in general, yet the barrier to entry is so artificially raised by school institutions and the FAA. College programs are the most approachable way for kids to get into aviation beyond those whose daddies can pay out of pocket for training. We'll continue to have a pilot shortage until young pilots aren't made to jump through a thousand hoops just to get their degree on the academic side *in addition to* flight school
It's ridiculous to me that stuff like VORs is something you're still required to spend time and money on and understand, yet they don't even exist anymore really and were mostly obsolete since the late 80s. But I digress lol
As an aircraft mechanic, I knew our it systems were fucked in regards to parts procurement, but never imagined other sectors of aviation to ba also fucked to such an extenr
Then you haven't learned sh*t... if you are a pilot you should have picked up on the fact that this dude doesn't know enough to know that he doesn't have a clue.
Well, this aged like wine
I was an IT professional at a bank. It frightens me that some of the processes that the bank relys on are still running in Cobol. That code is as old as I am and I'm retired.
You're probably older than the code unless you started coding when you were born.
@@knoahbody69 The only Cobol code I ever wrote was as a part of a course while I was in university. Even then, Cobol was considered to be old, obsolete and strange. No, I did not write any Cobol for any client ever.
Modern versions of cobol are actually not bad to write in. The problem is crust that has built up over 40 or 50 years.
If only AC DC had written a song called COBOLed instead of Snowballed 😉
I work for one of the top 3 banks in Europe and I literally worked on a Java batch application that is older than me just last week. I'm 29.
I develop IT infrastructure topology to companies. And, o boy, what I see in the backbones of IT infrastructure is insane sometimes. Many companies don't have a budget constraint in infrastructure, specially when they are the basis to the business or when they have a cultural bias towards innovation (like tech companies and startups in general). But some companies are properly insane. I recently run a metrics analysis in a BANK, and what came back was... wow. People running HR, PD and salaries through Excel sheets, with no backup... IT personal making adm backup into their personal Google Drive accounts... Memory and latency beyond critical performance with 7 years old hardware with no support or spare parts... Crazy after crazy
Honestly though, I see this more as a problem in the IT world than non-IT. Upgrading an old vehicle to a newer one is much more straightforward than upgrading to a new software solution. Software never works as intended or as expected and workers find ways around that, but not even good training can overcome the experience with the old system. It's like giving out a new vehicle with the steering reversed. Sure, on paper, just steer the other direction. But of course everyone would crash.
"If builders built buildings the way programmers wrote programs, then the first woodpecker that came along would destroy civilization." -- Weinberg's 2nd law
I worked in small business as part of an MSP. So many HIPAA violations. No backups. 15 year old virus ridden servers. It was disgusting. Never use your credit card at gas stations, Dentist offices, lawyer offices, even hospitals.
this is hilarious because this is exactly my experience... when your office workspace still uses a more primitive version of a spreadsheet you know whats up and I have no incentive to advise the change so... why bother... -_-
infact it got so bad that I want to try ask airlines how the heck they still get parts for such PCs cause finding them is getting rarer in my part of the world
in short that yes I want to adapt what airlines use for the house... its THAT BAD
tl;dr the office still uses DOS applications to run... and I get laughed at posting comments at LTT if they have answers... well LTT community being LTT community you get the idea
hate to break it to you, but that is common with banks. They don't have IT as part of the base mandatory infrastructure because technically one person with a spreadsheet can run a single bank...just not process more than a few customers per shift ;)
This "Good Enough" mentality with accountability only to the share holders is honestly one of the most pressing problems we have today, period. Not just in the airline industry, but in industry altogether. It's a primary driving factor in the income gap eroding the functional middle class, driving a portion of global inflation and encouraging corporate corruption on a massive level.
Yes! Everyone has bought into the "we can fix all our problems by cutting red tape" mantra that politicians use to pretend their budgets are going to balance because of those magic red tape savings, and everyone seems to have forgotten that this is PRECISELY what regulation is supposed to be for. Making sure that the thing the free market is optimising for is actually the outcome we want, because they will default to whatever happens to be in shareholder interests and sometimes this is in the public interest and sometimes it isn't.
It is just much harder to actually do the work, think it through, and implement something rather than just make noises about evil "red tape". (And if we are honest, how many voters are going to recognise that work as being valuable?)
If the FAA funded a new fast, modern, extensible system and then denied licenses/permits to every airline that didn't update to support it, airlines would fall over themselves to implement it and not lose the revenue. Problem solved.
Thats not the "good enough" mentality, but the "why not be greedier" one
so f the shareholders and make all industry state owned for the public
@@ibn_klingschor If every worker was staffed in the Army and fleet, we'd have full employment - and nothing to eat
@@the_expidition427 uh how. Armies must be well fed to fight in wars
as someone who works in IT, you'd be surprised how much critical infrastructure runs on ancient, fragile code written many decades ago but hidden through a nice, modern interface
not at all
Yeah, I work as a building automation specialist and it's crazy all my clients thar still do that. The only industry that I contracted with that never was like that was pharmaceuticals.
That's some John-Wick-level stuff
Doesn't the entirety of the world's banking system still run on COBOL? Or am I misremembering?
@@grzegorzbrzeczyszczykiewic563pretty much yeah
The fact that airlines don't yet extort every passenger individually is the only good thing in this story
And that's only because of their complacency. Certainly not owing to any sort of benevolence.
There’s no such command in the GDS
I don't have concrete evidence, but I suspect hotel and car rental companies do this by tracking browser cookies and IP addresses. Sometimes I have secured better rates when using a VPN via another state or country.
And don't always believe the "Last room left at this rate" flags.
@@simongchadwick Fun trick for hotel booking. Check your booking price, and if its to high for you, reset your browsers cookies and go back to the webpage. (Do not reload, input the original link again.)
On certain booking sites, this is enough to have the algorithm generate a new price.
That's what competition is for -- they can't extort too many passengers or they lose business.
As an airline pilot, I can confirm that this is all true.
Our scheduling system still runs on Windows XP. An early version of it. I’m not kidding.
Also, regarding the experienced rebooking agents part of this. Also very true. Just a few months ago as I was about to fly a flight home where my sister was a passenger, I realized it was going to cancel. 2 hours before it canceled in LGA, I knew there was a flight home out of JFK at a slightly later time and was able to get my sister on the train to JFK so she made it home. Just knew that because when you do it day in and day out, you eventually figure out what goes where and when.
You mean the windows XP that is not supported anymore and is vulnerable to hacking?
IS this ssue w a US problem or worldwide?
@@hassanabdulaziz7275 I’m not entirely sure, honestly.
@@hassanabdulaziz7275 Worldwide. Ask any air crew and they've used a horribly outdated software at least once, probably still does. There's probably only one good EFB software in the market right now and even that is not great
@@knoahbody69 Hopefully it's on an isolated network where it can only talk to other machines that run Windows XP.
The “reverse tragedy of the commons” you described is basically the prisoner’s dilemma, mutual cooperation is beneficial on aggregate but individual non-cooperation is locally better
I think a better word for the "reverse tragedy of the commons" is capitalism. Why would individual non-cooperation be locally better? Competition is a disease preached to us by the mutually cooperating owner class.
@@CC3GROUNDZERO Go starve in Venezuela and then tell me how well the alternative works. Oh wait, you wouldn't be able to because even if you survived you aren't allowed to complain.
Great video. As a former IT professional I've used the economic uncertainty of today to try out some other fields besides tech. It's crazy how you can tell management that a planned upgrade/fix will cost so much less than if the system explodes, but they never listen. Then of course when things do explode, they try to throw you under the bus. I've always made sure to save emails of this happening, and it's saved my ass before. I hope things change but I sadly doubt they will.
but why are IT systems not designed to be easily upgraded or transferred to new systems. seems to me that should be the main requirement but it's always so complicated and takes great resources.
@ron black The best way I can describe it is to imagine it like changing the gauge of rail track. A larger gauge functions very similarly in concept to a smaller gauge, but you still have to replace the rail itself and the wheels on train cars. Then, there'd be a transition to the new system where a loss of productivity occurs until everything adjusts to the new system. It's pricey in the short-term but not as pricey in the long-term as overloading a smaller system until it fails. Short-term profits in capitalism still prefer to keep the smaller gauge and hope for the best. In IT terms, businesses are using old code/hardware and hoping that it holds instead of transitioning to new systems with more capabilities. Code would need to be optimized for multi-core, hyperthreading capable processors, and parameters would need to be able to accommodate the massive amounts of data an upgrade could provide. Things get complicated quick when you're dealing with trillions of ones and zeroes that need to be processed then sent somewhere else
@@ronblack7870 Well, you can't switch to new systems that don't exist. You need to create those.
TBF a lot of IT professionals do also want to upgrade things that are working fine and not about to explode. We had to pull teeth to get an expensive ($50k) bit of network testing automation equipment, but we could have still done the tests with less automation, it would just be more work for us. The equipment was a big switch that let us set up test scenarios automatically instead of having to hard-wire them.
@@ronblack7870 Designing for upgrades or transferring to new systems is an expense that most businesses are unwilling to pay.
I'm a computer scientist from germany and I knew the guy who came up with a similar system for German Rail. Whenever he heard about a delay of a train he stood there for a while frozen calculating the delays impact for the whole network.
And btw the key to solving NP-hard problems is using heuristics. Otherwise, your system will only be good up to an engineered limit and will reliably fail beyond that.
Understand too that Europe's national rail operators often operate different & incompatible ticketing systems, & people desire them to be more standardised & integrated like GDS is for aviation, so as to facilitate international rail travel
I’ve been told that the dutch rails still run in DEC/Compaq/HPe/VSI OpenVMS on virtual alphaservers
My friend is studying aviation (basically pilot degree) and they asked me if I wanted to see their flight computer. As someone studying computer science, I find computers very interesting. They pull out a paper with a paper wheel pinned on and hand it to me. I was so confused.
Don’t bring the E6B into this
Different thing all together. Pilots carry E6B in case of electrical failure, where an analog computer may be all they have to do certain calculations. Technically it’s a flight computer, at least that’s what it’s called, but it’s more like a flight calculator.
@@ducky8075 smart duck
I trust an E6B more than I trust anything else. It's what I learned on. In the event of a total electrical failure you will be returning to your E6B, lap board, and pencil. Total reliability.
As a computer science student, you will eventually learn that "having the latest tech" is not only not achievable, but also dangerous. Robust and failsafe systems are preferred for critical areas, and in case of electrical failure pen and paper are king
I used to work for an airline using Amadeus. Unfortunately they started going with a graphical skin on top of Amadeus and forbidding us to use the command prompts we were using. That meant that all that changed was that now we were only allowed to click on buttons, which added significant loading time per click. Worse, the system didn't actually change, it just limited what the agent was allowed to do. So once it crashed or broke (which happened all the time) all you could do was wait until the system sorted itself out. Glad I left that job, so much frustration for nothing
Are you by any chance talking about AlteaCM?
@@greecoboost it was not :)
I work at a company selling a GUI rivaling the one from Amadeus and some customers are so happy to switch to ours because it's so much more simple and responsive
@@samconstantinou2335 if you're allowed to just use amadeus you'll be fine haha. Very clunky but once I got used to all the command prompts I was so so fast
It was GoNow?
As an IT worker, this story hurts my soul. These problems are all SO solveable.
And also very common
from a software standpoint they are solvable. The real problems are economics and cooperation as was also mentioned here
Manager's response: "solveable? I don't think so, there's not ticket and no budget for that - go working on something useful like new UI design for customers so it will look nice, but be more confusing for the user".
@@proosee Far too clear and far too few buzzwords to be a real quote from a corporate manager. XD
@@proosee of course their is money, it's just a matter of if they want to. All they need to do is have these billion dollar companies all put a bit of money on the side to fund a new or existing industry group to put in the work. Lots of industries have some done similar things. for example the banking industry has SWIFT.
I worked at a major international airline during a period where it realised its' old 1960s reservations system was no longer fit for purpose. The options were a) build a new modern system from scratch b) apply duck tape and bolt some clever little fix programs on top of the existing system. The airline chose b) because senior management was (probably rightly) terrified by the risks associated with a) - development effort, time and cost / amount of testing required / smooth cutover / integration with other legacy systems / PR damage due to inevitable problems. When your business is almost completely dependent on IT systems, reluctance to change is understandable.
The failure rate of replacing the old airline systems is astronomical. So they have instead focused on building additional functionality outside of the old GDS systems to supplement the old limitations. But it is all very complex and spending money is no guarantee of success.
Yeah, but then you do this for 60 years and no one is now able to even just duck tape the crumbling mess, much less reconstruct it in modern technology (especially since the documentation is lost to time if it even existed in the first place)
Legacy stacking will NEVER go away as established companies/ industries such as the airlines and telcos aren't willing to spend Billions on developing and implementing new and improved systems. Biggest problem is the further you kick the can down the road, the more complex and expensive it gets. Sabre was developed in the 1960's and is the main reservation system still used today.
European passenger rail companies looking at the situation, after having done the exact same thing and now refusing to go back to interoperability in booking systems: finally, a worthy opponent!
weren't they given an ultimatum by the EU though? Where they have to collectively present a plan on an interoperable booking system, otherwise the EU will make their own and the companies have to use that one.
EDIT: granted, that would still mean it is at least a while before it is used.
@@zephyros256 they were given an ultimatum, but I sure haven't heard anything of them even bothering to do anything about it, lol (if anything, some were actually lobbying for the EU to abandon the idea). And if the EU decides to make something of their own, we may well never see it in our lifetimes.
You could have formatted this comment in a non zoomer ticktock algorithm way.
"Me typing out an entire comment with context, sentences and punctuation: Im gonna ruin this mans whole carrer!"
@@kepler656 who fukin cares
@@bjoerkengard smaller discount train operators are disrupting rail travel in Europe right now cuz of EU.
The really interesting part is you can kind of apply this stuff to basically any industry that was big enough to get computers going in the early days. Banking, Hotels, Medical, Airlines, Power, Shipping, ect.
This is an exceptional video. Growing up, both of my parents were flight attendants, and helping them navigate the command line interfaces of their respective airlines is what got me into computer science.
One other tidbit - their bidding process was absolutely INSANE by modern standards. When flight attendants have the appropriate level of seniority (seniority is what EVERYTHING is based on in the industry), they have the ability to pick and choose which flights they want to work. Back in the day, this was done by distributing a gigantic packet to every flight attendant which contained every flight itinerary for that airline for the entire month. When I say gigantic, I mean it. The bidding package was often 600+ double-sided pages, in tiny font, with almost no white space. The amount of paper they went through was absolutely ludicrous.
My dad's best friend when I was a kid was a United pilot and his wife was a flight attendant. I remember them comparing notes because my dad was a truck driver and he bid runs as well. The airline stuff was absolutely crazy. The upside of it was when my dad was gone on the truck and Dale was in town I used to get to go to Stapleton and hang out with him once in a while.
I used to be a travel agent for a large travel agency based in NYC and had to create itineraries and fares by hand using commands in Sabre to generate tickets. I even used validation plates and had to hand write them as well depending on the air carrier.
We have those as well when I was with KLM. Whenever I use that validating machine to issue an MCO, the passengers ar the counter are always startled with the loud smashing sound it makes. 😂
@@kingthranduil8807 That's hysterical.
What is a validation plate?
@@vishalsyoutube A validation plate is a metal embossed plate with the airline name and logo along with their three-digit IATA issued accounting code. This allowed you to validate an airline ticket on that airline or one in which they had an interline agreement with.
@@rustydogcfl thanks!
The other issue with replacing and updating the GDS is that there are other industries that also rely on it.
Hospitality, car rentals, some entertainment/tourist stuff, and even some bussing and cruise ship systems.
There have been a couple different pushes to update the GDS by big players over the last 20 years and it never moved past the idea stage.
One of my previous jobs I worked at one of the main Hospitality companies, we updated and rebuilt our systems every 10-15 years but they still hooked into and used the underlying 1960's GDS.
I can tell you from experience in an unnamed payroll company that changing from a command line system to a graphical one for the ticket agents will NOT make things smoother, faster or easier. When you know how to run your command line with the amount of information needed you can fly through the process. Make that a point and click and the delays between even just screen refreshes will turn a 5 minute process into 20 minutes.
Yes!!! Good on you.
True, but there's still no reason why you couldn't invest in improving the system - making the process of getting an re-booking more flexible, scheduling including crew information, making it harder to do a mistake - or make the mistake less hard to fix afterwards.
From my experience I want to echo what many here are saying: IT might be expensive, but the one of the things even more expensive is to wait until it crashes and everything burns. How much did SW have to pay for their Christmas chaos? You can probably save up to 90% of that by improving your system before it crashes in the worst possible situation. Down time is bad, but down time during Christmas is worse.
But as the IT guy, you don't get to hire the people needed to fix the system, you're just there to keep it running, day after day, while yourself most clearly seeing these things will - one day - crash and burn an you're like Cassandra forced to have no one believe you, or no one care.
Very much this, it's really important to get it right.
A lot of the improvement comes in training new employees. Text systems are deceptively fast for experienced employees, but painfully slow to adapt for the new.
@@Steven_Olson that's why TUIs (text user interfaces), which are a hybrid of CLI and graphical, are amazing. Powerusers can rip through using hotkeys while noobs can point and click their way through menus.
The decrepit IT isn't just on the airline industry. I work on the UK rail system, and we similarly have a back end system for train schedules (called TRUST) which is the backbone of the timetable information supplied to passengers and the public. There are other systems too for things like crew scheduling and rolling stock planning, but these often don't talking to each other, so when a major route gets closed for an unplanned period of time due to an infrastructure issue, or emergency, that's why it all falls apart so dramatically, affecting services often far away from the area where the problem is.
"There are other systems too for things like crew scheduling and rolling stock planning, but these often don't talking to each other, so when a major route gets closed for an unplanned period of time due to an infrastructure issue, or emergency, that's why it all falls apart so dramatically, affecting services often far away from the area where the problem is."
Indeed. I think the video poster is trying to imply that we should try to replace aviation with public transit trains and that would solve it, but public transit does the same kind of stuff. And it doesn't have the speed that we need in North America and Australia. It can't take 20 hours to go from LA to NYC. Five or six hours is already too long.
Good news is that Lockheed and NASA have developed a fuselage shape that lets you break the sound barrier with a barely audible sonic boom, so hopefully we'll get supersonic over land again....so it will be more like 2 or 3 hours to go from LA to NYC. 🙂
As an operations/aerospace engineer inside Boeing, I am always impressed by how concise and entertaining your videos are! This one take the cake and yet its just the tip of the iceberg in terms of the constraining factors that GDS has within the part 121 world.
What giant industry or corporation these days *doesn't* settle for good enough? I've seen it happen over and over. Businesses have to be good to get big, but once they get big, they stop trying to be good and start trying to stay big, and being good and staying big are sadly different directions for most businesses.
IBM, Goggle, Facebook, PayPal, airlines, manufacturing, railroads, trucking... list goes on and on. It's like the old saw of running from a bear... I don't have to be the best in the industry. I just need to be marginally better than you in some manner.
To be fair to airlines, they are alot more regulated, and most of their stuff has to first get approved by the FAA before they can use it, and that's expensive and slow.
7:54 - this sure brings back a lot of memories accessing my former airline’s reservation system to check bookings and flight availability. Had to have a cheat sheet to remember all the codes needed!
As for Southwest not allowing booking via GDS - it’s a legacy of the early days of the GDS. Back then, those GDS were owned by the airlines (for example American Airlines owned Sabre, while United owns Apollo which was merged into Amadeus). These airlines, in a way of trying to squeeze their competitors, chose to charge other airlines for every booking made on their GDS. Even worse, some of the airlines, namely American, dicked with the algorithm of the GDS so that their flights get first display on the GDS because they found that travel agencies back then would often offer customers the first flight offered because they were used with the fastest flight being first in their OAG routing book. In one instance, they dicked the system so that flights from New York Air, a competitor, was listed last in the system. Why? Because they dared to challenge American on the New York-Detroit flight.
All that led to Southwest to say, “F*** you, we just won’t sell on your asswipe system” And that’s why they’re not on a GDS today.
I miss the OAG. If *only* us 'normal customers' could have access to what was in that book back in yesteryear!
Southwest uses Amadeus now and for the past several years at least at the frontline (airports).
All true
I'm watching this after the Crowd Stike Crash 😂
Ι am working for the reservations department of an airline and although relatively newbie and a young person, I heavily dislike the web format of our GDS (Amadeus ARD). I was trained in it, but I surely prefer the old-school cryptic commands. In reality, it is a lot simpler and faster to work this way. Instead of clicking "Reprice" then choose the passengers, hit "Confirm" and then "Update Fares", I just write FXF and there it is, plain and simple. There are also many things that can only be done with cryptic commands, the scenarios that you have mentioned are actually a piece of cake. Anyways, that was a very nice video and it is always nice to see your hard work being appreciated :) Truth is that I was heavily inspired by early Wendover videos and decided that the commercial side of airline operations is equally important and interesting to the flight operations one. So thank you Sam!
So true
Cryptic ftw, until you have to automate it.
My buddy is a flight attendant with another major, so take this with a little grain of salt I guess, but they said a major component of the Southwest meltdown was that when flight crew call in to get new assignments they call into the same number that everyone else does… which, when you cancel 10,000 flights can be a little hard to get through on!
🤨
Problems will also arise for old software programs that have been written in old programming languages that are no longer part of modern IT education. Whenever the existing generation of programmers reaches pension age, there is nobody to replace them. This becomes a problem whenever the old systems need to be adjusted to operational changes.
I work for a bank. We have many young cobol devs.. it is bewildering. Were'd they learn it??
So at Delta we use that software but internally we call it DLTerm. So let's say you want to look up a flight from say Atlanta to MSP (Minneapolis Airport) you would write on the system SATLMSP then it will spit out in bunch writing the flight time, departure time and different classes available in letters it's pretty archaic system but it's one even us devs don't want to touch because if you fix one thing something else breaks and most of the people who originally built these systems retired back in 2020. Fun fact it's quite common to leave something broken because if you fix that something else might break or their might be a cascading effect so if the thing broken isn't important or critical we just let it be.
So new systems should be (and should have been) implemented in stages instead of all at once, is what you're saying.
@@andyjay729 correct
And I thought the 240p, 3 inch, 24 color displays from 1989 in eco were the most ancient thing in the industry…
You would not belive how much code that society is depending on is legacy code from 80-s or myb 70-s .. ( similar situation is inside all banks and credit cards ) ..
And yes I have huge respect to software developers from that period ..
@@danicic87 Make that 1960s for both banking and the core of the GDS. COBOL for the win, all credit to the amazing developers and the IT staff who chewing gum and rubber band to keep it working
@@ariandeelsii sorry, you're giving too much credit to these developers and IT. I know quite a few "old timers". IT was very simple back then: call support, which was 99% of the time IBM. 20 years ago I had a teacher who complained that "now you have to do all yourself, in my day IBM was great, they even came without us even calling them, because the mainframe had an error and called IBM by itself".
as for developers, we're talking about an extremely conservative industry with rules set in stone for decades. those systems are extremely simple (and because they're simple, they're so reliable). New developers have to work with the latest-and-greatest always pushing what's possible because that's what the market is about. Have to be inventing solutions with unclear and ever changing requirements.
a coworker of mine, her dad worked at the bank in the 70s. he did the software in COBOL for the bank. the guy was a construction worker who grabbed a COBOL book one day and showed up at the bank when they were hiring.
@@ariandeelsii yep, that's what we used in 2019 for USAA Federal Savings Bank
@@KingLarbear I've been part of a bank upgrade before, it's criminal the rickety-ass tech our entire economy is based on.
You might owe them an apology in light of resent events.
Exactly the same problem in banking, and for the same reason - early adopters of computers with enormous legacy systems.
Military is the same. US ICBM launch control is all in OS360 assembler.
@@amcluesent That is kind of the point nowadays though... Low sophistication means low susceptibility to cyber warfare, which is what you need with ICBMs. As long as they guarantee a counter launch before the enemy nukes hit the silos, that is good enough.
Same with insurance industry...
@@Spectification I've heard this a lot and while I think it has merit I feel like it's more an accidental benefit then a planed strategy.
@@eldermoose7938 For sure its not a planned benefit. Its just a the biggest argument why not to upgrade.
Fantastic simplification. After 17years in Travel Tech can’t tell you how many GDS integrations were cancelled due to ridiculous GDS developer licenses, etc. Amadeus, Sabre, et al made it too expensive to integrate so companies just do direct integrations with key sales channels. As you say, the APIs are old and severely limited, when combined with extortionate integration fees, it quickly becomes not worth integrating…
This aged perfectly
Even those pictures you showed of the British 90’s travel agent IT kit was still just a front for another ancient it system that was running on even more ancient kit in a dusty data centre somewhere near slough
Ah, but what happens when "good enough" is no longer true? I think that's where we are now. It is PAINFUL to travel now. I've traveled for 40 years now and the past four years I have regularly canceled flights, which never happened before. The terrible TSA lines, delayed flights, charging extra for a cup of water and carry-on luggage, the overbooked flights, crew not available, crew is late, paying extra to SIT NEXT TO YOUR CHILD, gate agents are the worst people to exist on planet Earth, the 90 minutes on hold with support, and terrible online booking experiences is too much. It's too much. I don't know if it's possible to incentivize better behavior, but I'm in favor of treating this like a crisis where the government helps get everyone on the same track. I mean, we literally had to write a law on how long airlines could detain us on the tarmac because airlines didn't care.
Air travel is practically dehumanizing at this point. I live 900 miles from my mother and we’d rather do two days of driving to visit her than a two hour flight.
@@abbysc417 This is why the US needs a high-speed rail network, it would free up the airports of all the small local planes that need to connect town that already have a rail connection.
@@rafski123 the us needs a Shinkansen bullet train system like japan
@@HalbeargameZ Shinkansen could be an option. Siemens' ICE trains have also proven reliable in many countries (Germany, Austria, Netherlands, Russia, China, to name a few). Also, there is the Alstom TGV, which is comfortable and good for high capacity with its double deck options.
One of the problems is that the current market actively rewards the airline that can get you from point a to b with the lowest initial ticket price. Stuff that was included in the ticket like seat reservation, luggage, carry-on etc, is now an upcharge because it looks more attractive on booking sites to have a low initial price rather than a fare where everything is bundled in. One of the reasons why everything else gets cut or worse in quality.
Time for a sequel to this video 😅
I appreciate that Sam marked Midway Airport in Chicago instead of O'Hare, as Southwest only flies out of Midway. This attention to detail is what makes Wendover so great and reputable.
Actually, Southwest *does* fly out of O'Hare now. They added it back in February 2021. However, it still flies a lot fewer flights to and from O'Hare than it does to and from Midway.
Southwest flies out of O'Hare now too. I know because the last time I flew there were only about 20 million advertisements for it.
Actually just flew in last Monday. Southwest was present.
Apparently, that's wrong, so Wendover isn't so great and reputable.
Another thing that contributed to the SWA meltdown was the ramp issue in DEN. A manager sent an email to all the rampers requiring everyone work overtime over Christmas. The message ended by saying if you didn’t show up you were saying you quit. I heard 90% of the DEN rampers called off sick. While not the full cause of the meltdown it made things much, much worse for no reason. A lot of SWA employees complain that the days of Herb are long gone and that the company no longer treats its employees like people, more like commodities.
That happens to every company after the founder dies.
Human ressources
"...the company no longer treats its employees like people, more like commodities"
Welcome to neo-liberalism.
That brings into play the other issue that airlines currently face, being understaffed. If they had more employees, they would be able to put more staff on standby for when issues arise. With current staffing levels, a blip in one place can cause the whole system to breakdown.
@@LutraLovegoodprotects company, not people. Hence the name. Manages a resource.
Anyone here in 2024 talking about Crowdstrike?
That's what I came here for too 🤣
Ayyyeeee 😂
No mainframes affected 😊.
Can I double like this?
This has to be one of the most amazing insider videos I've ever watched. I work for one of the largest retailers in America and we run on Mainframes. As this story explains, the Mainframe didn't crash and that wasn't the issue. The issue, or at least one of the main issues, was that the scheduling system only scheduled planes, but not the crews. And the rule on crews timing out wasn't accounted for anywhere, which made things even worse.
I have been working in IT for 10 years, and I honestly do not remember what Mainframe means.
Modern mainframes are remarkable pieces of technology. It's a shame they get written off as outdated or legacy or the source of the problem most of the time.
@@dankdank1891 Mainframes are centralized servers, usually built by IBM nowadays (I believe Hitachi also makes modern mainframes but I know little about it). Mainframes are much more powerful than decentralized servers (the servers people usually think of when they say "server") and usually have greater reliability than decentralized servers. Some mainframes configurations have been running for decades without downtime, thanks to redundant systems and sysplex. Linus Tech Tips (surprisingly) has a pretty good video about mainframe hardware
Many of these crew "time out" regulations likely did not exist when the systems were built; so they weren't updated to keep up with changing laws.
@@dankdank1891 A mainframes sole purpose, including having a fully customized z/OS or a very stripped down version of Linux is to process transactions and store the data as quickly as possible. It is all about efficiency. While the hardware has gotten a little better, the software hasn't changed much, why rewrite code that was already optimized to almost the fullest extent decades ago. Not every company needs a mainframe, only those that deal with large volume of concurrent transactions.
Rental car companies are like this as well, or at least they were 10 years ago when I worked at Enterprise. The in-city locations had Windows 98 level GUIs, but the airport locations ran on command line programs. And it was a constant rotation of employees, so there was usually only 1 or 2 people who had that senior-level experience because they were permanently based at the airport.
“Command line” is missing the point - CLIs can actually be very efficient (as also the video remarks). The main problem is the backends behind them.
...And the lack of education so younger employees can't use the proper interface and need to use the graphical layer which can't actually make use of all the functionality.
I work for a legacy brand airline and it’s amazing how old the technology is, just like you showed.
But also like you show those of us who have worked for a few years know how to get ourselves around the system, the way regular passengers can’t.
I'm working for an airliner as a developer and we use Amadeus. We are moving to NDC and while better it causes a lot of headache and takes ages. Don't get fooled by the "new" in NDC because it is based on XML and feels already outdated before it's implemented fully.
Really nice video, very well explained! Btw, 16:11 takes away a common misconception that we are dynamically changing the prices or based on individual customers. There are only a certain amount of slots for a price so if they are all sold the price increase, not because you are looking for a second time to the same flight.
yuck XML. 🤮
As they sat in congress, re: the price increases, YOU LIE
What's this new-fangled JSON stuff, right? Now try explaining to Google they shouldn't pin TLS certificates ... Everybody doing NDC is doing at least something wrong!
@@mdnpascual it's even worse than it sounds.
i was part of a dev team which planned the implementation of NDC and it really is unbelievably stupid
@@petergamache5368 thing is they are extremly slow to build ne standards like NDC because so many Airlines have a say in it that it overcomplicates the whole process.
And because of that, NDC is the next best thing.
Worked for KLM customer service, the onboarding alone was 2-3 months just because the booking system with those codes was so outdated and difficult it took a lot of practice. I still have nightmares of codes like AD15OCTAMSCDG and shit like that. We had to do that manually while we had customers on the phone to order tickets as well. Boy was that a shit show. From my group of 15 new hires 11 quit (incl me) after the 3 months onboarding period, and the outdated and annoying system was a contributing factor for all of us and one of the root issues for lack of motivation, high pressure and even stress for some.
I would much rather enter AD15OCTAMSCDG and have instant response with accurate availability than fill in endless boxes on a slow fiddly online form and wait for pages to load and see half the picture .... Just speaking as a Amadeus native user dealing with corporate travellers... I guarantee that making bookings, changing bookings, re-issues etc take a fraction of the time compared to NDC where , when forced to use , we grind to a halt. But yes, I've had my laughs with KLM call centres ... "married segments" HAHAHA
Only lasted 8 months on that posting. Typing out entries can be confusing at times but when you got to talk to a passenger on the line it gets more confusing.
_I'd rather work on the ramp servicing aircraft than go back to Airline call centers again._
As someone who, on the daily, works with Amadeus reservations and Altea CM, it baffled me how easy things became after a while. At first it seemed like everything had its own entry and to some extent it’s true, everything ticket agents do has a backbone to the basics of creating a PNR.
It's not difficult, you're just not suited to it. Hundreds of thousands of travel professionals use it every day. It's fast, it's accurate and it's extremely powerful. Airlines have been trying to build GUIs and alternative systems to replace Native for over two decades now, and nothing comes even close to replacing it.
Southwest Airlines laughing at this video after surviving Crowdstrike with their Windows 3.1 systems
A good presentation, and ultimately the key phrase is 'what's good for the stockholder ...'. Every corporation will end up cutting it's own throat because it can't see past the quarterly earnings that are 'good for the stockholder'. It truly amazes me that their stockholders, while earning a little less, will earn far more in the long run if they take the long view where public good and futures investment is considered.
B-b-b-but, and hear me out here...
Money n o w
/s
Absolutely agree, absolutely hate the way corporations work these days
Not here and now? What, you think HIGH FREQUENCY TRADING is a joke or something??
🔫🤡 🔫🤡 🔫🤡 🔫🤡 🔫🤡
@@TheDoomerBlox 60 fps 46 khz trading
@@heartache5742 The remaining 2 Khz were lost in the bufferbloat, unfortunate reality.
it's about the ceo execs and boardmembers not stockholders. If a ceo comes onboard a profitable well run company..cuts salaries and benefits, slashes the maintenance and puts its departments on starvation budgets profit jumps short term as the company feeds off itself. 5 years later his reighn has been very profittable, he the board and the execs have recieved massive bonuses..he leaves and the next ceo takes charge of a broken down barely functional organism in it's death throws and the profit and stock prices start to reflect that
This reminds me what happened with internet in early 2000s. Initially we built many interoperable protocols for emails, chats, file transfer, voice calls that were software agnostic - you could be using one software program on one end and a different one on the other end and they worked well together. It worked remarkably well in 90s. Today they all except email died out and been replaced with non-inteoperable corporate cloud software as service products - facebook, google drive, onedrive, etc. It was interesting to learn that in airlines industry this olden approach survived for so long. In nutshell the problem with interoperable protocols is that once they get adopted it becomes too hard to change anything about them to the point where they get replaced by private closed propriatery product that is just many miles better.
And even email is slowly being replaced by Exchange.
@@Yggdrasil42 Isn't Exchange a brand of email server software?
I feel like the only semi interoperable software I have is blender for the single reason it has a bunch of file formats it can save as or load except for whatever the he’ll hbaked is
Usually not even replaced by better protocols, but by worse software with a lot of market share behind it
If you know about raising money, #1 money is infinite
#2 only if the problem is big and obvious enough is it worth spending
#3 when you do spend, make sure it's super good and novel so it lasts a long time.
So it makes sense these are the patterns in airlines.
I was a consultant for an airline two decades ago. Their IT systems were 30 years old, allowing only four simultaneous connections for some important things. Although it must be said that the bunker for the old mainframes was big and awesome.
I used to train agents on Worldspan 20 years ago. Worldspan was basically TWA and Pan Am's PARS system and replaced Delta's Datas II. It is exactly the same now as it was 20 years ago. So a system written for TWA and Pan Am is still in used along with SABRE (which is even older). That should instill confidence
Yes, considering that Pan Am shut down in 1991.
These systems were developed in the 60’s so they’re now 60+ years old. Unfortunately the ‘new’ NDC (XML) interfaces need to be backwards compatible so while interfaces may have changed, the underlying logic/limitations have not.
This is one of the single MOST IMPORTANT VIDEOS for travelers to watch. Helping them (and myself) understand why booking anything but a highly-typical itinerary can be a headache and seemingly extremely expensive. Thank you Wendover for making this video 🎉🎉🎉
The most widely ignored software design rule is to "fail gracefully" Then the rule becomes "Lessons not learned, will be repeated "
The most widely ignore software design rule is to fail *early* - better a small, localized crash that can be fixed by the developers actually responsible and while they remember all the context, than keeping it limping along with a bunch of squelched warnings and eventually cause a chain reaction of other failures.
@@leftaroundabout so "fail early, gracefully (as in, still return error but don't corrupt anything else), and learn from that error."
And then Borrow at the Bank of Bodge because it needs to be fixed yesterday and the customer is screaming at your clueless manager.
make the system fault-tolerant, not fool-proof
Honestly?
It's best to just focus on fault recreation.
Never ever try to 'handle' anything that isn't an exception generated by an unreliable system (connection timeouts, retry, etc)
Number one priority, record what happened (public entry point) with which parameters, the stack trace and every version number required.
Ensure your build process stamps unique version numbers
write a new unit test to catch the fault before deployment, patch the code and turn the handle on the release process.
Don't have one of those?
God help your damned soul, because I wouldn't touch your company with an infinity barge pole.
Very good summation of the Status Quo. I liked your definition of the current airlines situation as "Reverse tragedy of the Commons".
There are two solutions: one is the mandated government minimum standard of operation; two is setting a voluntary inter-airlines independent commission who sets a standard that every airline is obliged to implement. Servers and applications have come down in price since the IBM System/360 was running that command-line interface (don't ask how I know).
Thank you for the video. Pretty much every aviation channel on UA-cam has made an oven-ready video on Southwest meltdown, but your Wendover video is the only one which had detailed the mechanism of the failure and the relations between the various software programs (apps?). I agree, the communication protocol may deserve the ½ century upgrade.
Greetings from the UK,
Anthony
So, in other words, SouthWest suffers from what all big businesses suffer from: PROFIT FOR THE SHAREHOLDER over all else at all costs. Profit for the shareholder over safety. Profit for the shareholder over employee pay. Profit for the shareholder over customer comfort. Profit for the shareholder over all else, always.
How convenient for timing. I went on delta airlines yesterday and mid flight, their entertainment systems(which run on linux) had to reboot because of malfunctions
Fun fact, Linux is the de facto OS for Supercomputers and Financial trading servers so the OS is not the limiting factor for why Delta's entertainment system had to reboot on your flight. Delta could easily make their entertainment system more robust so that it could handle software failures and hardware malfunctions while providing 100% uptime availability. But they will never do that as it's an order of magnitude more expensive to do so and they have no financial or PR incentives to spend that kind of money.
This problem of not adapting to newer software technology solutions is not just limited to the Airlines. You could also find it in a other legacy businesses. Take for example, Neiman Marcus, an old-school retailer. When I used to work for them right before the Pandemic, they were still using command line software developed in the 90s for their employees to clock in and for their inventory and logistics. Even though the computers were Relatively modern looking desktops. For example, we would open up one of the proprietary programs clicking the icon on a windows desktop, and the program’s window would pop up displaying a black background with green text like it was 1983.
Pretty good video. GDS are still using cryptic entries, but Amadeus is not any longer cryptic native since it migrated out of TPF, it is just backward compatible as travel agents are actually way faster using cryptic entries.
And yes GDS only provides interlining routes as possible routes, this is however a commercial limitation rather than a technical limitation. Basically your itinerary is under the same e-ticket, and whoever is the airline in the e-ticket is the one in charge of bringing you to destination.
But yes passenger could take a combination of separate tickets for the same itinerary. And several companies saw this as an opportunity to create "virtual interlinings". Comes with some drawbacks though: you have to pick-up you luggage and re-do the check-in mid way, and you are on your own if the connection is missed. "Virtual interlinings" companies try to mitigate those drawbacks with agreement with the airports for your luggage, and with insurance in case of missed connections.
The #1 asset of an airline is their safety record, the #2 is their reliability record. You would have to be crazy to buy tickets on SWA the week of Christmas 🎄
Not really
no, the #1 asset of an airline is its frequent flyer program
No, it's proximity to where people live and low costs. Ryanair is a universally hated - not disliked, no, outright hated - airline despite its 100% safety record, in over 25 years of operation they never had a single lethal accident. KLM, which was involved in the Tenerife disaster, the most horriffic air crash in history, is a lot better liked. The reason people fly Ryanair instead of KLM is that you can fly from easy to reach local airports. Within an hour of travel by car, 3 Ryanair bases are located from where I live: Eindhoven, Düsseldorf, and Weeze. The closest international operations base of KLM is Schiphol, which is twice as far away.
My experiences with KLM are rare but were very positive. I still fly with these cunts from Ryanair because Eindhoven is just too close and too easy to travel to (even by train and bus it's just an hour away from here) to justify travelling twice as far and paying a slightly (or significantly) higher fare just to avoid them.
As a software engineer, this whole topic just breaks my heart…
UA-cam algorithm is crazy showing me when so many airlines couldn't operate due to an IT outage
The old model I learned in college still remains true; “To error is human, to really muck things up requires a computer.” 😂
This was by far one of the best, most informative presentations you’ve made to date. And I watch every single one.
I was seriously schooled. And I worked for United Express doing baggage service from 1998 to 2000.
Great video, I was under no illusion that the IT systems in the airline industry were old and complex, but it was neat to learn specifics. Love this sort of "behind the scenes look"
This video seems extremely well-researched and presented. Even higher than the other Wendover production videos. Good job
I was sitting in the airport during the FAA system restart. It was really weird to get notifications that our flight was delayed right before the news broadcast behind us said all flights were grounded. No one knew what to do or whether their connections would be there when they landed. Even the flight crews were like "we're going to board and try to take off at 9:30, but no promises we'll get off the ground"
The US needs to privatize its air traffic control service. After Canada privatized its service, it became of the best if not the best air traffic control service in the world.
@@CharlesWT-TX you clearly didn't watch the video. Shareholder value is the root cause for technological debt of airlines.
Ancient IT just saved Southwest from Crowdstrike. It's hilarious but points out the value of heterogenous IT systems because when everything runs the same OS then some vendor pushes an update a fecalstorm ensues.
Retired from a large software company 2 years ago after 32 years. When I was first hired in 1990, there was then-antiquated airline software that was associated with a product I worked with. It was a real chore to find skills to mend anything that went awry, or to scaffold a system in order to test fixes to the archaic code. I believe that the software is still used at many airlines, and it looks about as friendly as the 3270 green screens in this vid. Memory was really expensive in the days when the software was created, hence the brevity of presented data. Amazing that such old code works as well as it does.
This has to be my all time favorite channel, the level of explanation in depth without getting too technical is phenomenal. Keep up the great work.
I used to work for a travel agent in the UK from 2000 - 2004 using Amadeus. I can't believe it's still being used in 2023!
Another great explainer video! As a veteran of the travel industry with over 30 years experience, I'm glad you packaged this conundrum in such an easily consumable way. And as a technology partner to airlines, we really hope to break down some of these inhibiting stalemate factors. Nice one, Wendover!
CLI are quicker a lot of the time than GUIs, at least for trained employees.
I did 2 years as a shop assistant and we used a CLI, boss pressed 8 buttons and could pull up any order. It was insane how shortcut keys worked.
you know shortcuts and GUIs are not mutually exclusive?
@@alexalcalaortiz368 no, but I can do Git commands quicker in terminal than via shortcut keys in an IDE
A) you could argue that the much increased training and adoption time may outweigh the benefits of a CLI, if the GUI is actually well designed, and decently speedy
B) the two should not be exclusive. I still hate how strategy games figured out decades ago that you can have an easy to use GUI and a million shortcuts at the same time, and still make it work (with, let's be honest, minimal dev time overhead), and a big chunk of commercial software still doesn't let me Tab through options, or hit Enter for OK on a pop-up.
A bunch of the ick here is using a CLI built decades ago for trained humans, as the machine-focused API. "Easy for humans to read" rarely translates into "easy for machines to read", leading to brittle code that's shoved in a box and nobody dares touch it ever again. A better design is that the CLI is simply one more interface over the same machine-friendly API, which can be designed to accomodate new concepts without risking breaking every existing client.
CLI versus GUI is a much falser dihoctomy than is usually realised, and I suspect frequently turned into a console war of character cell terminals from the fucking 1970s, versus actually admitting you're using a bitmapped display to emulate that terminal. Modern twists are possible, with results displayed graphically (eg, numerous RDBMS frontends), and maybe even using graphical rendering to call out things like the boundaries of values rather than text.
And even then, the problem isn't terminal vs bitmap. It's lazily developed--out of laziness, cheapness, or being in a hurry--software that uses bottom-tier design, without having things like commands as a concept in themselves, or reflection that can be used to automatically derive a scripting interface, or even being able to visualise data in a fairly raw way.
I use voice commands at work now, it's much faster than manual inputs
Wondering what could’ve possibly led to this being in my recommended on July 19th, 2024 🤔
Telecom had the same problem: backbone of most of the systems was written in Cobol, & you could typically count the number of people on staff that knew Cobol on one hand. When those people retired & something's not working correctly, the hunt was on to find one of the remaining people...to save money, telecom management would "kick the can down the road" rather than fund modernization (too costly), with predictable results.
There's a few other places where this "if only you knew how bad things are" situation is an everyday reality, banks for example with legacy cobol & fortran systems with more modern solution on top or to the side of the older stuff. Utility like electricity & water in some countries are like this too.
Cobol is working better with floating numbers(if talking about banking), also features hardware independence
I haven't heard of Fortran in a very long time.
I can't tell you how long I've wanted to know this and how much I enjoyed finally learning about it!! That being said, this shit is so complicated I could barely keep up with it and I have no clue how you were able to grasp it enough to make the video itself. Back in the early 2000's I called the airport to look into a flight that someone had booked for me, I was lucky enough to talk to someone who genuinely cared enough to want to share some useful info with me. Again it's irrelevant now, but at the time it was priceless. She gave me a number to call if I wanted to book last minute flight via any available cancellations. That was basically like being given the most valuable thing ever in the realm of booking flights. It's a shame that I was never able to make use of it, but it meant something to me at the time.
Two things of note 7min in. We don't call reposition flights "passengerless deadheads". Deadheading is a term we'll use for crew about exclusively. We'll call them reposition ferry flights in the industry. Also, the maps image of the FAA headquarters with the ground stop that was issued wouldn't have happened at that location. That would have been at the Air Traffic Control System Command Center (ATCSCC) in Warrenton, VA. It's right next to Patomac TRACON.
This was one of the best explanations of legacy technology and the problems it causes. I've worked in enterprise for my entire career and it was initially shocking (and still terrifying) to see how much critical technology is stood up on tooling created so long ago. They're all hard problems to solve for sure, but this kind of problem is going to happen more and more across industries.
It's so funny to me that Southwest escaped with almost no issue with the Crowdstrike crash BECAUSE they have this old software still...
as an employee for an airline mentioned in this video - they have tried to replace these command line tools with fancy new front ends, but we always go back to using the good old green and black screens as they just work so quickly! Long live command line interfaces!
Blame the Common Use workstation standard ... which still says pointing devices are optional.
Yes, finally, I work at a travel agency that's specializes in the marine sector, I honestly from the bottom of my heart to the depths of my soul HATE THE FRONTENDS give me my cryptic line and I will book you in 30 seconds to wherever you want, sure, for the newcomers the cryptic entries are a pain to learn BUT once you master them you cannot in good conscience use the frontends.
Let's say that I use sprk to book from ATH to FRA, it LITERALLY takes me 2 minutes to create the reservation, on Sabre I need no more than 30 to 40 seconds plus I have all my quick keys programmed... NDC is just a fancy way of jamming more ancillary services to the pnrs it does nothing to help with the reservations... Actually it's even slower since for each reservation the GDS must download all the information that it's not needed...
Agreed, when you get used to command line interfaces and know how to get the best out of them, it's super fast
@@nicknellas1287 One of the main problems with modern software development is that non-technical people (who are often in charge of the software development) have trouble understanding anything that's not a graphical UI.
I imagine that this explains a lot of the useless new frontends: they are easy to sell because the CEO can touch and feel them, even though they don't fix any underlying problem with the ancient backend systems.
Worked retail in 2000. Had green screen look up for inventory, price, wholesale price, reorder etc, way more than a sales clerk would need. I only knew a couple commands, but it was damn useful and fast. If it said it had it, we had it. Then they replaced it with heavily locked down Windows running on comically underpowered hardware, as in typing would lag behind significantly. Everything took forever and told you nothing. Insult to injury the colour bit depth was set to 8 causing eye bleeding dithering that, if switched to 24bit manually via some workarounds, ran at least 10 times faster making it useable and looked as it should.
I knew as this company started cancelling flights that it had to be a lack of upkeep for tech, damn shame for all of those missed flights but I sure hope they fix it soon.