Re:Move deliberately, fix things: How Coinbase is building a...
by starpilot on Sunday May 20, 2018 @04:24AM
(#17111801)
That practice is as old as engineering itself, but computer programmers are just starting to learn it.
[ Parent
| on HN ]
Re:Move deliberately, fix things: How Coinbase is building a...
by hyperbovine on Sunday May 20, 2018 @07:39AM
(#17112316)
Isnt it hilarious to see this being touted as some new and insightful way of developing a product? CS undergrads should be forced to learn about how bridges, skyscrapers, and airliners get built.
Re:Move deliberately, fix things: How Coinbase is building a...
by zrobotics on Sunday May 20, 2018 @11:47AM
(#17113245)
I actually think the curriculum works directly against this. I know my homework assignments tended to be much more unmaintainable than even a one-off python script that I would hack together. It might be interesting to design a class where, instead of weekly homework, students are required to continually maintain/modify the same source over a semester. Weekly projects that will never be touched again actively discourage writing maintainable code, and I don't think that is a good lesson to be teaching.
Re:Move deliberately, fix things: How Coinbase is building a...
by PeterisP on Sunday May 20, 2018 @03:15PM
(#17114305)
I recall reading about a teaching system where homework involved taking submissions of previous year's students and adding features on top of that, providing students with all kinds of practical insights on what makes code easy or hard to maintain.
Re:Move deliberately, fix things: How Coinbase is building a...
by AlexCoventry on Sunday May 20, 2018 @12:12PM
(#17113406)
They do often learn about disasters like the Therac-25. https://en.wikipedia.org/wiki/Therac-25
Re:Move deliberately, fix things: How Coinbase is building a...
by AlexCoventry on Sunday May 20, 2018 @12:09PM
(#17113392)
It depends on the context. Facebook's cavalier, exploratory attitude mostly worked because they were building an entertainment. Software development tends to be a lot more cautious in domains like finance, where errors can cause serious, actionable damage.