71–80 of 106 posts
So how do other airlines track where their planes are? Radio? ADS-B? How about their crews? In-house mobile-phone app? The Southwest telephone-based system sounds super low-tech. AirTags might work better.
Everyone here is at least a little misinformed regarding their software and it's simulation abilities. They have control systems with the data they need, and this data is fed into the flight scheduling & monitoring UI. Their tech works, it just gets bashed on because it's different than a lot of other airlines. Used properly it can account for these things. What causes things like this is overbooking. Overbooking hap…
I consulted for a few large carriers and this was my conclusion also. In fact most very large American enterprises have all the same issues. With airlines the problems simply become really visible all at once really quickly when things go wrong.
For the folks here on HN whose experience is mainly in Silicon Valley: it's hard to appreciate how little the execs at these companies care about software. They don't care, not even a little bit, and they definitely don't care what your opinions are. Their priority is growth, the stock price, and answering to the board (not necessarily in that order). The only carrier I heard about that cared for their IT was Continental and they were bought by United.
Compounding the issue is how staffing works at these companies. Being a full-time employee at an airline is considered attractive because you get benefits including cheap airline tickets. My understanding is that the airline ticket perk used to be much more awesome than it is today.
In any case, airlines bend over backwards not to have full-time employees. Depending on the company they have a vast army of contractors (on shore and off shore) who are on a revolving door policy lasting from 6 to 18 months. These folks come in, get trained up for a few months, turn around tickets in a grueling and dehumanizing environment, then they get to take a hike for a while until they can come back on another short-term stint. I had a colleague who had been full-time at SWA, he said his job literally only consisted of training people up on the systems, he rarely wrote much code himself, he was there to 'keep it all together.'
But honestly, the current crisis is not a surprise. The IT systems at big American enterprises are truly horrific. It is decades of homegrown software "integrated" with decades of acquisitions where systems are smashed together on short timelines in service of quarterly goals.
If you want to find the true culprit here, look up at the broader structure of the economic system. This mess is created by how we run the economy.
Earlier quoted context omitted.
Ah so bean counting treated IT as a cost center and cut CapEx to the bone. Running things with no reserved capacity, crew wise and system wise, is just asking for a catastrophic fall off the cliff style shutdown. Edit: a simple queue theory can explain why no reserved capacity leads to system breakdown.
The question is, will they still come out ahead even accounting for the catastrophic shutdown? Maybe once a year is fine
Earlier quoted context omitted.
Yup. For me they had brand prominence because of various customer friendly things like bags fly free, sit where you want, etc. It will take a mountain of good reports from other people over a significant period of time for me to want to fly with them.
This is the first time I’ve ever seen the plane-boarding hunger games of sit-where-you-like referred to as a customer-friendly thing.
Basically what I read was the former CEO was accountant educated and focused on stock price and financials and ignored system upgrades for like a decade. That coupled with their decentralized nature of operating (no hubs or star pattern) the recent storms caused their systems to not “know” where people or planes were even though the people knew where they were. That knowledge is the critical feature because hours of…
Not sure it's as cut and dry as "CFO becomes CEO and overly focuses on financials" in Southwest's case. In 2021, Gary Kelly (former CFO/CEO) announced his retirement and picked Bob Jordan (CS trained, having previously worked on the AirTran integration) to take over as CEO, which seems a realization that Southwest has some serious technical debt to address. All businesses have highest priorities at a given point in t…
Earlier quoted context omitted.
Apparently Southwest is one of the biggest offenders in the airline industry when it comes to not investing in their IT infrastructure. You can only get away with that sort of debt for so long.
If that's true, and I have no reason to doubt it, the blame will fall squarely on the team who were asking for IT investment, and not the leadership who said no to them.
https://www.theregister.com/2022/12/28/southwest_outdated_it...
Earlier quoted context omitted.
Apparently Southwest is one of the biggest offenders in the airline industry when it comes to not investing in their IT infrastructure. You can only get away with that sort of debt for so long.
If that's true, and I have no reason to doubt it, the blame will fall squarely on the team who were asking for IT investment, and not the leadership who said no to them.
Earlier quoted context omitted.
Airline systems are some of the oldest computing architectures in use. The Sabre reservation system was created by IBM for American Airlines in the 1950s — when a computer filled an entire building floor and was built out of vacuum tubes — and remains actively used today by multiple airlines. The programming language has been switched several times over the past 60+ years, but essential compatibility remains.
I can't remember where I read this and I can't find a cite right away, but I bet someone can confirm-or-correct me: the 6-letter reservation code for your flight? That used to be an explicit pointer of memory in oldold Sabre.
Earlier quoted context omitted.
Airline systems are some of the oldest computing architectures in use. The Sabre reservation system was created by IBM for American Airlines in the 1950s — when a computer filled an entire building floor and was built out of vacuum tubes — and remains actively used today by multiple airlines. The programming language has been switched several times over the past 60+ years, but essential compatibility remains.
I can't remember where I read this and I can't find a cite right away, but I bet someone can confirm-or-correct me: the 6-letter reservation code for your flight? That used to be an explicit pointer of memory in oldold Sabre.
Some observations, though.
- Via the departure control, manifests, etc...they DID know where their crews went. Any employee flying non-rev, deadhead, or on duty is recorded by employee number. There was likely a problem with keeping that synced with the crew management platform.
- Large parts of this have little to do with technology. Point to point versus hub and spoke means less slack. And recently, SWA has removed even more slack trying to drive revenue with aggressively optimistic schedules, overbooking, tight crew slack, etc. If they didn't pull pack the schedule in small, manageable pieces like the other airlines did, there's no system that would save them. Seems likely they didn't pull back enough before the storm. There's a brink you can't go past, and a rate of cancels you shouldn't exceed. You have to guess right earlier.
- Crew unions have historically negotiated out anything that looks like big brother, location tracking for example...even on-demand "I'm here". So lots of things that could have made this better didn't exist , on purpose.
- A fair amount of the reset is just practicalities of using planes empty of pax to get rested crews staged in the right cities and bags to the right places.