Open source’s “Greatest Hits” go global

When the GitHub Archive Program sent 21TB of repository data to Svalbard last summer, the Arctic Code Vault became the flagship of the effort to preserve open source software for a thousand years. That storage run involved 186 reels of piqlFilm, a digital photosensitive archival film, and marked the start of a broader preservation strategy rather than the end of one. Today, GitHub is extending that work by distributing what it calls “Greatest Hits” collections—the most popular and depended-upon open source repositories on the platform—to four additional locations around the world.

Each Greatest Hits set includes the core collection of roughly 17,000 repositories printed on two reels of piqlFilm. That count reflects the reality of modern software development: with 99% of new software projects depending on open source code, these repositories form the foundation of much of today’s technology. Alongside the most depended-upon projects, each set also contains a random sample of about 5,000 additional repositories, intended to give future historians a better sense of the breadth of software being built by developers today.

The distribution follows the LOCKSS (Lots Of Copies Keeps Stuff Safe) principle, meaning the same data now lives in several independent institutions. The four sets have been donated to three of the world’s key knowledge preservation bodies, with the fourth kept at GitHub headquarters:

  • Bodleian Library at Oxford University, England — One of Europe’s oldest libraries, whose holdings already include codices over a thousand years old.
  • Bibliotheca Alexandrina, Egypt — The modern successor to the legendary Library of Alexandria and a partner of the Internet Archive.
  • Stanford Libraries, California — Located in the heart of Silicon Valley, Stanford has a long track record in digital archiving.
  • GitHub headquarters — The fourth set will be preserved and displayed on site.

Each donation comes in a bespoke art case designed to hold the hardened microfilm reels, the same type used in the Arctic Code Vault.

Beauty as a preservation strategy

The decision to house the reels in custom enclosures wasn’t just about aesthetics. At a January advisory board meeting, GitHub’s advisors stressed that beauty is a practical requirement for long-term preservation. The logic, as Alexander Rose, executive director of the Long Now Foundation, put it: “If you don’t make it beautiful, it’s for sure doomed.” Humans are more likely to care for and retain objects they find visually appealing and that carry perceived value.

GitHub commissioned artist and engineer Alex Maki-Jokela to build four museum-quality cases that mix traditional aesthetics with 3D printing and AI-generated artwork. Maki-Jokela describes the work as aiming to honor the spirit of open source software and the generations of science and engineering that underpin it. He notes that the software and printing techniques used to create the cases sit close to the edge of current capabilities, while the symbolic content is meant to reference both the engineering advances of the past century and the older, more timeless stories that will remain relevant even as technology’s forms change. Maki-Jokela has documented the process on Medium.

For the partner institutions, the program aligns directly with their core missions. Richard Ovenden, the Bodley’s Librarian, framed the collaboration as a natural fit: libraries and archivists are as much the advance-guards of the future as the custodians of the past, and the open development practices of the software community mirror the open approaches that libraries have long championed in scholarly communication. Dr. Mostafa El Feki, Director of Bibliotheca Alexandrina, pointed to his institution’s existing work in web archiving, book digitization, and 3D artifact digitization as preparation for taking on software preservation. Michael Keller, Vice Provost and University Librarian at Stanford, noted that the program’s reliance on multiple holders of the same data echoes Stanford’s own LOCKSS-driven philosophy.