So Zipline uses drones to deliver critical and lifesaving products precisely where and when they are needed, safely and reliably, every day, across multiple countries.
They’ve done it more than 30,000 times already.
How does Julia help?
Guidance, Navigation, and Controls Engineer Tucker McClure explains:
“Our legacy system was built in C++ around vehicle dynamics created in Simulink, where we used code generation tools from MathWorks to turn the Simulink model into C, and we incorporated that C. It’s difficult to work with and excessively difficult to update, and it simply can’t do a bunch of the things we need to do.
“The new Julia simulation already has a variety of advantages, with essentially the only disadvantage being that a bunch of our software team is more familiar with Python.
“We are using Julia to develop our new aircraft simulation platform, and we chose it for the mathematical expressiveness of the language, performance, and general flexibility. The core of our team is meant to spend their time writing mathematical models that need to run quickly and in a variety of different ways.
“If Julia had never existed, we'd be doing a lot of metaprogramming: writing Python that writes C++ that's actually as fast as we need. We've all done it, but it's pretty clumsy.
“Mostly, we're focused on enabling flexibility, and that enables us to do new kinds of testing and analysis, and we simply wouldn't have done that without a flexible language, because it would take longer and would therefore be at a lower utility.”
What does Tucker say is the best thing about Julia?
“It's so much fun! Julia really enables people to race forward with their ideas, and that means people can be more creative and happier.
“We are working to turn the single aircraft simulation into a set of tools for all kinds of analysis we couldn't realistically do before. And we're in the middle of a major transformation of our internal systems making Julia a core part of how our company operates.”