an air transport parable
Feb. 17th, 2025 09:46 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
totientMirrored from Bluesky.
Seventy years ago, at the very dawn of the jet age, American Airlines got so tired of taking airline reservations on index cards that they did an absolutely audacious thing: they decided to write a computer program to do it.
In those days this was the sort of thing that only militaries did, and well-funded militaries at that. It cost them fifty million dollars. Those are 1957, gold-standard dollars, enough of them to buy ten or a dozen jet airliners.
A few other airlines also thought of writing computer programs but American's was the best and pretty soon most other airlines were using it. It grew new features, most of which its customers couldn't live without once they had them.
A few decades passed and the system got larger and clunkier. It did an awful lot of things, none of them especially well. And because the underlying technology was so ancient, it was brittle and hard to update and expensive to keep running. But too much depended on it, and so it kept on going.
About twenty years ago, a few very very smart people had enough hubris to think they could replace it. "How hard could this be?", they asked. Ten or a dozen of them sat down to start writing code, expecting it to take a month or two.
A year later there were 250 people working on the project.
Two years and another fifty million dollars later, the project was canceled, having failed to produce a usable airline reservation system.
At least this time they were 2007 dollars, not 1957 dollars.
(this is not a story about airline reservation systems)
Seventy years ago, at the very dawn of the jet age, American Airlines got so tired of taking airline reservations on index cards that they did an absolutely audacious thing: they decided to write a computer program to do it.
In those days this was the sort of thing that only militaries did, and well-funded militaries at that. It cost them fifty million dollars. Those are 1957, gold-standard dollars, enough of them to buy ten or a dozen jet airliners.
A few other airlines also thought of writing computer programs but American's was the best and pretty soon most other airlines were using it. It grew new features, most of which its customers couldn't live without once they had them.
A few decades passed and the system got larger and clunkier. It did an awful lot of things, none of them especially well. And because the underlying technology was so ancient, it was brittle and hard to update and expensive to keep running. But too much depended on it, and so it kept on going.
About twenty years ago, a few very very smart people had enough hubris to think they could replace it. "How hard could this be?", they asked. Ten or a dozen of them sat down to start writing code, expecting it to take a month or two.
A year later there were 250 people working on the project.
Two years and another fifty million dollars later, the project was canceled, having failed to produce a usable airline reservation system.
At least this time they were 2007 dollars, not 1957 dollars.
(this is not a story about airline reservation systems)
no subject
Date: 2025-02-18 06:02 pm (UTC)(the above is about reservation systems, but everything you look at has more crucial moving parts than you can see.)