The code survived.
The reason didn't.
One that nearly went
Free software argued itself into existence in public, on Usenet, in the 1980s: the announcements, the objections, the replies. Usenet had no archive. Messages expired off the servers that carried them. When Google assembled the largest surviving collection in 2001, a large part of the years 1981 to 1991 came from magnetic tapes that Henry Spencer had kept at the University of Toronto, for no better reason than that he had not thrown them out.
The code from those years is everywhere. The reasons survived on one shelf.
That is the shape of the loss, every time. Code gets copied. The mailing list, the argument, the bug report, the note explaining why the fork happened: those get tidied.
This is what FOSS Archives is for.
There is almost nothing here yet. One holding. That is why this is the right time to decide how it should work.
The rule is not complicated. Three copies is storage. Many independent copies, held by people who do not answer to each other, is preservation. So the archive does not ask you to trust it. It asks you to hold something.
The GNU Manifesto
Richard Stallman, 1985. First printed in Dr. Dobb's Journal, March 1985. About 4,800 words. The document that explained, before the code existed, why an operating system was going to be written and given away.
It carries its own permission: anyone may make or distribute verbatim copies, in any medium, provided the copyright and permission notices are kept. So you may hold it, and pass it on.
It currently exists in these independent places, which we checked on 29 August 2026:
- The Free Software Foundation, the current texthttps://www.gnu.org/gnu/manifesto.html
- The Internet Archive's Wayback Machine, the same page as captured on 26 January 1998https://web.archive.org/web/19980126190329/http://www.gnu.org/gnu/manifesto.html
- The Internet Archive, a scan of Dr. Dobb's Journal, March 1985, where it was first printedhttps://archive.org/details/1985-03-dr-dobbs-journal
- This site, as a plain-text file with a README and a checksum/holdings/gnu-manifesto-1985/gnu-manifesto.txt
gnu-manifesto-1985.zip: gnu-manifesto.txt, README.txt, SHA256SUMS
You now have the same bytes we do.
sha256 of gnu-manifesto.txt
0d228059c6f6b79c9ec5b4c96305f65cbd88e35ee1fa7013c57b117400df9cb9
To check it, now or in 2050: sha256sum -c SHA256SUMS. The README says what the file is, where it came from, and why you may keep it.
If this site disappears tomorrow, that copy does not.
Good.
Copies taken from here since 29 August 2026: …
How it will work
No single custodian. Anything held here is meant to be held elsewhere too, by people who do not answer to us.
Deletion leaves a mark. If something is removed, the record that it was here stays.
The context, not only the code. Mailing lists, bug reports, release notes, the argument that led to the fork.
Open formats. Plain text where plain text will do. Nothing that needs a particular company's software to open.
Fixity is verified, and the hashes are published, so that anyone can check a copy against any other.
Nothing to renew. No subscription, no login, no account that lapses.
Still looking for
A second, independent scan of Dr. Dobb's Journal, March 1985.
Any of the 1985 replies to the Manifesto, from Usenet or from print, held somewhere other than the collections above.
If you have either, or know who does: hello@worthbuilding.org
Holdings: 1. Independent copies we have checked: 3. Copies taken from here: ….