Comment by osullivj | Hacker News Reader

Comment by osullivj | Hacker News Reader

Wonderful piece for those interested in early WWII aviation. There are lessons for software development in there. Brad Cox's classic Planning the Software Industrial Revolution [1] highlights the importance of gauges to Eli Whitney's project to built firearms with interchangeable parts, and asserts we need a similar technology in software to enable construction from reusable components. In Most Dangerous Enemy [2], Stephen Bungay points out that it took 2.5 times more man hours to build a Spitfire than an Me109. German production engineering was far better than British. This piece illustrates that perfectly. Rolls Merlin engine was a great piece of design engineering, but it was inefficient in terms of production engineering. What is intriguing about this article is the detail on the effort needed for Packard to port the Merlin design to their production engineering standards, and then for Rolls and Merlin to share design changes. Effectively sending each other pull requests!

[1] http://bat8.inria.fr/~lang/hotlist/free/licence/papers/cox/C... [2] https://www.amazon.co.uk/Most-Dangerous-Enemy-History-Britai...

> need a similar technology in software to enable construction from reusable components

People have been attempting this for a long time with only moderate success. It's where object orientation came from: objects were supposed to be reusable. Similarly with UNIX pipelines. The new enthusiasm for package managers produces some absurd results (left-pad) but is generally heading in the right direction, as is the trend that everything that's popular enough becomes a language feature or in the standard library.

Yeah, object orientation never really worked for making reusable lego-bricks.

The functional programming people have been at this game for far longer (and are building on top of mathematics started around the 1940s). There's some encouraging developments, including some that are making it into mainstream languages.

Yes, libraries (including standard libraries) do tend to work really well, too.

Interesting, given that I've read that Americans were far better at mass production than the Germans. My impression is that Germans engineered their hardware better than Americans but they were not as good at mass producing it fast.

So if I understand correctly, the British was exceptionally slow at manufacturing compared to the Americans then.

It depends upon your definition of "slow". At mass production, maybe. But maybe faster at changing and iterating on designs due to being more flexible and adaptable.

I read years ago that the Spitfire required twice as many labor hours to produce than the Me-109. The Germans were pretty good at simplifying the design so it could be mass-produced.

Not always German tanks where complex prone to breakdowns and took much longer to build.

Reply on news.ycombinator.com