The Tower of Babel and the Fighter Plane (2013) [pdf] | Hacker News Reader

The Tower of Babel and the Fighter Plane (2013) [pdf]

melconway.com

Great read!

Complexity is evil evil evil (my take on it: http://www.embeddedrelated.com/showarticle/132.php)

Incidentally, today’s automobiles share this problem of a massive amount of on-board code and the bugginess that accompanies it. A few years ago, some Mercedes owners found that when they pushed a certain button on the navigation system the driver’s seat moved.

</quote>

I have this sinking feeling that my death is going to occur because of some interaction between software/firmware design and an unexpected situation (e.g. stuck in an elevator between floors because the firmware is expecting internet packets from Twitter or something). Hope I'm wrong, but the complexity of electronics seems to be increasing fast enough for system reliability to decrease.

I doubt that scenario you mentioned will cause your death. The firemen will get you out.

Yeah, it'll probably be a car crash or a hospital equipment failure :/

This is the Mel Conway of "Conway's Law" named by Fred Brooks in The Mythical Man Month. Don't skip it just because it seems to be about the F-35; it isn't. It's about complexity, concurrency (the risk thereof) and---at the very end---about abstractions for fitting huge systems into your head.

Seconding this recommendation. It's not bashing the F-35 as an aircraft, but taking aim at the program. Sadly, as he points out, it's happened plenty of times before. The F-35 program just happens to be egregious because it started out so big.

Highly recommend this read.

An interesting problem that came up during development of the Navy version, which has to land on carriers: when the landing gear and hooks were first tested, the placement of the back wheels was such that they created a sine wave in the arresting cable which just happened to be "down" when the hook moved over the cable, meaning the hook wouldn't catch and the plane could never land this way. They then had to go back and redesign the arrangement of the landing gear and hook.

The key to making strong systems is to make each component do one thing, and do it well, the Unix way. Each plane does one thing, does it the best. When the leader deploys each plane to do that job, he can count on it. The last time we tried to make one plane do it all we got the F-111, which also didn't do anything the best, except the FB-111 made a fine bomber. Past, repeat, etc. Don't build the complexity into a component; build it into the overall system, the whole computer, not the individual plane or program.

I'm having trouble accessing this - quick copy-paste link of Google cache for anyone else with the same trouble: http://webcache.googleusercontent.com/search?q=cache:WYm78Yt...

Weird: strong start, leading up to a tantalizing claim, and then ... it just ended.

Reply on news.ycombinator.com