Open Source Is the Quiet Engine Behind NASA’s Latest Deep-Space Instruments

When the James Webb Space Telescope (JWST) finally left the ground last month, it marked the culmination of more than two decades of engineering. Less visible than the golden mirrors and cryogenic instruments is the software stack that will turn raw photons into the images and spectra that make headlines. Much of that stack is open source, built by a global community of researchers who often have no direct affiliation with NASA.

That relationship did not emerge overnight. NASA has long shared its data openly—anyone can pull images from the archives and reuse them without restriction—but sharing code is a more recent shift. For missions like JWST, which has been in development for over 20 years, open source only entered the picture in the last five years or so. The change has been driven in part by the expectations of the scientific community and in part by policy shifts that have made open development the new normal for government agencies.

From Astrochemistry to the Space Telescope Science Institute

Arfon Smith, a Director of Product Management at GitHub, has watched this evolution from both sides. His PhD at the University of Nottingham was in astrochemistry, studying the structure of the interstellar medium—the material from which stars and planets form. It was during that period, before GitHub existed, that he began writing software and recognizing its role in scientific work, from individual analysis scripts to making sense of the vast datasets produced by missions like the Mars Reconnaissance Orbiter.

He later joined GitHub as a science evangelist, working to support researchers on the platform and connect them with relevant open source communities. Among the initiatives he helped catalyze was citation support, which lets academics get credit for their software contributions. Smith then left to run the data archives at the Space Telescope Science Institute (STScI), overseeing data management and data science for Hubble, Kepler, TESS, and JWST. That role gave him a direct hand in how software shaped NASA's flagship missions.

How JWST Actually Uses Open Source

Python is the de facto language of modern astronomy, and the scientific Python ecosystem has exploded in popularity over the past decade. Central to that shift is Astropy, a community-coordinated collection of tools for astronomy. Missions like JWST do not exist in isolation—they build on a substrate of libraries and tools that predate them.

For JWST specifically, open source software powers nearly every layer of the science workflow:

Notably, this is a different model from earlier flagship missions. Hubble's software was largely developed behind closed doors. The JWST effort, by contrast, happens in the open on GitHub under the Space Telescope organization, allowing astronomers around the world to contribute to the tools they will ultimately rely on.

The Policy Shift Toward Open Science

Government agencies historically hesitated to work in the open, but that stance has softened considerably in the past decade. NASA's recently announced Transform to Open Science (TOPS) initiative builds on a series of policy changes around open code and open data. Beyond NASA, groups like 18F within the GSA and the USDS have pushed for more open, technology-driven approaches across the US government.

Space agencies have also established public presences on GitHub:

This openness creates direct opportunities for researchers, students, and independent developers to study and build on the software agencies are creating.

A Journal Built on Pull Requests

Smith's own major open source contribution is the Journal of Open Source Software (JOSS), which he founded in 2016. The journal treats software publication as essentially an academic pull request: authors submit their code, reviewers examine it in GitHub Issues, and once it passes peer review, the software gets a formal citation. JOSS is open to all fields of research and has published over 1,500 papers in its first five years, with roughly 70 volunteer editors.

Some papers that went through JOSS have since become highly cited in space science, including Gala, a Python package for galactic dynamics, healpy, a toolkit for spherical data analysis, and dustmaps, a Python interface for interstellar dust maps.

Where to Get Involved

For developers interested in contributing to space science, the entry points are accessible and numerous:

Astronomy is often an observational science, and the same applies to joining these communities—start by following repositories, reading issues, and understanding how the tools are built.

What JWST Might See

Predicting what JWST will find is nearly impossible—that is the nature of frontier science. But its infrared capabilities guarantee certain kinds of observations will be transformative. By operating in the infrared, the telescope can peer through dusty regions of space where stars and planets are forming, resolving those processes in detail that was previously unattainable. It will also look back toward the earliest phases of the Universe, roughly 300 million years after the Big Bang, when the first stars and galaxies were igniting.

Exoplanet research is another likely source of headlines, as the telescope's infrared sensitivity is well suited to characterizing the atmospheres of planets orbiting stars outside our solar system. The discoveries will come, but the code that enables them is already in the open, waiting for contributions.