|GitHub's Arctic Archive|
|Written by Sue Gee|
|Saturday, 16 November 2019|
The most bizarre announcement made at GitHub Universe 2019 was that GitHub is planning to store every open source GitHub repo in disused mine on the Svalbard, a remote island closer to the North Pole than to the Arctic Circle.
The goal of the GitHub Arctic Code Vault is to protect the world's software in the event of an Apocalypse and to preserve our modern way of life. Six thousand of the most popular GitHub repos, including Android and Linux, have already been stored and on February 2, 2020, GitHub will capture a snapshot of every active public repository to be deposited alongside it
Introducing the video clip relating to this seemingly bizarre initiative, GitHub CEO Nat Friedman explains as human progress depends on open source software, the code on GitHub matters and it should be preserved for posterity, concluding:
And so at GitHub we decided we should do something. And to be honest we might have gotten a little carried away ...
In his preamble Friedman reflects on how engineering knowledge of classical Rome had to be rediscovered in order to build the dome of the Duomo di Firenze. Eager to win a contest to design this glorious structure, Filippo Brunelleschi unearthed ancient texts from Vitruvius and others.
While Friedman's example is taken from history, there are perhaps better parallels from fiction. Even though GitHub's choice Svalbard is at the opposite extreme to the Utah Desert the scheme puts me in mind of the sci-fi novel "A Canticle for Leibowitz" where in the aftermath of nuclear holocaust a religious order is devoted to preserving for posterity the knowledge that had been contained in books that were burnt, not only in the catastrophe but by those who blamed books for civilization's downfall. When the ancient texts known as the Memorabilia are discovered centuries later deciphering them is an almost insurmountable problem. In the case of the code stored in the GitHub Archive it will include technical guides to QR decoding, file formats, character encodings, and other critical metadata so that the raw data can be converted back into source code for use by others in the future.
In view of climate change, Svalbard, which is already home to the Global Seed Vault, seems a reasonable choice for long-term storage and the data is first transferred to 3,500-foot reels of film that uses silver halides on polyester designed to survive for a millennium.
The Arctic Code Vault is just part of the GitHub Archive Program which aims to provide a range of storage solutions with three "pace layers" characterized in terms of heat. Hot is updated near real-time; Warm monthly to yearly and Cold every 5+ years. The other principle is "LOCKSS" - Lot Of Copies Keeps Stuff Safe. The Hot and Warm data will be kept in multiple locations and duplicate film reels for the 100,00 most-starred and most-depended-upon in the Arctic Code Vault will by stored in the Bodleian Library at Oxford University. In addition the GitHub Archive Program is partnering with Microsoft's Project Silica to archive open source code by writing them into quartz glass platters using a femtosecond laser.
So at least there is method in what we might still consider madness.
or email your comment to: email@example.com
|Last Updated ( Sunday, 17 November 2019 )|