I’ve been reading science fiction since I learned to read, so I am very, very familiar with post-apocalyptic life in all its many forms. We’ve been imagining the “last people on Earth” for more than a century now, and I have to say that my own conclusion is “avoid this situation if at all possible.” Rebuilding everything from scratch would be very hard work, and not likely to succeed.
There have always been people who want to act out these fantasies right now, at least partially, for a little while. The basic back-to-the-land, dropping off the grid, simple life turns out to have a lot of permutations. Many people are motivated by religious or ideological ideas, expecting and hoping for the unsatisfactory present to collapse, leading to a new age of moral renewal.
Others are driven by extreme self-interest, looking for personal immortality even in the face of the expected collapse of the unsatisfactory present. Silicon Valley billionaires are particularly interested in life extension, pouring GDP levels of money into technologies that are intended to keep them alive for many current lifetimes, possibly “forever”. (Unsurprisingly, they are able to find people who will take their money as long as they are willing to spend it, possibly “forever”.)
And then there are technologists who have a bug in their ear about how transient our current technology actually is. Many of the great technical achievements of our civilization are invisible, and are already disappearing if not gone.
For instance, my own master’s project in the 80’s describes software written for an operating system that doesn’t exist anymore, in a language that doesn’t exist any more, running on computers that don’t exist, made by a company that doesn’t exist any more. (The paper itself was supposed to be preserved on microfilm and maybe paper, but looking for it right now I can’t find any trace. It probably still exists, but it’s certainly not readily accessible. Not that you or anyone needs to see it.)
A more important example is the US National Archives, which is required by law to preserve “historically significant” materials,.This certainly should include many aspects of the space program, such as the software from the Apollo moon landings. Software that was written for an operating system that doesn’t exist any more, etc., etc.
Most of us don’t worry too much about this. I mean, I’m just as happy that the world can’t see my youthful follies. I don’t take myself or my own work all that seriously.
But, some people are concerned that this short-term mentality is harmful. Indeed, not thinking that our actions will matter in a thousand or more years is a good way to, say, fill the atmosphere with CO2 and melt all the ice. Or carelessly wipe out all living things that, oops, make human life possible.
And, of course, some people are just romantically attached to the notion of creating something that will last a thousand or ten thousand years. Aside from the archivists’ reflexive desire to preserve a record of our civilization, there is a certain super villain-ish desire to have future beings know your works. (See Ryan North, Chapter 9.)
Can you say, “hubris”?
This fall I read about an interesting effort along these lines, The GitHub Arctic Code Vault. Conceptually, this is an effort to preserve a snapshot of the GitHub open source software to be readable in 1,000 years [2]. Physically, this is a steel vault containing images (on tape?) of QR codes that record source code from Github. The vault is located near the Borgen Seed Bank, geographically and psychologically.
A lot of thought has gone into the design. The outside is decorated to make it attractive, i.e., to encourage you to believe that it is valuable and to look inside [1]. (This is the inverse of the principle that radioactive waste stores should be repulsive—highly visible but not at all inviting.)
So what’s in it? It contains a snapshot of every github project active on 02/02/2020. This is millions of projects, from huge, important things like linux, to tiny, little used dinks. Github projects generally include source code, documentation, test data, and metadata about who did what.
This snapshot was apparently encoded in QR codes which are stored on long lasting tapes. The QR codes can be decoded by eye if necessary, and are relatively robust to minor damage like scratches or stains. I.e., you can reconstruct what’s on a QR code even if it’s a bit crumpled.
“On 02/02/2020 GitHub captured a snapshot of every active public repository.”
(From The Archive Project)
There is also documentation explaining how to read QR codes and what Github is. Apparently, there is also a small library of “how to” books. And a copy of wikipedia! And, interestingly, an almost completely random collection of fiction
Phew! Can you say, “hubris”?
Obviously, any effort to create this kind of time capsule must face long odds that it will survive at all, and if it survives, someone will notice it, and if noticed, will be interpreted in any meaningful way. Archaeology is full of lost artifacts, ignored artifacts, and artifacts that are not understood or comically misunderstood.
It’s not nice to root against them.
But I can’t help wondering if I even want them to “succeed”.
Of course, there are plenty of technical quibbles. QR codes may be ubiquitous today, but they weren’t around even 50 years ago, and probably won’t be in another fifty. And if you don’t use digital computers—because quantum computing—what does a “QR encoding” of data even mean?
I have to wonder about the art work that is supposed to be alluring. Artistic taste is notoriously fickle, so it’s got to be a crap shoot trying to appeal to a completely unknown audience in a completely unknown context.
But anyway.
I’ll also note that survival for 1,000 years generally doesn’t hinge on putting stuff in a single, very pretty box. That’s a formula for getting it looted. If you want something to last 1,000 years you want there to be many, many copies; stored in many, many media; with lot’s of cross references so people know that it exists and what to look for. (See, for example, the account of the rediscovery of Lucretius.)
For that matter, the entire idea we call “source code” has changed rapidly over the last 70 years, and surely will change in the near future. When I was a lad, “source code” was something that mostly came in big, heavy boxes of punched cards. Twenty years ago it came on portable disks. These days it comes from github. Who knows what it will look like inf the future?
Worse, the actually ideas expressed in what we call “source code” has evolved dramatically. When I was a lad, we were moving tiny bits of data around tiny little computers. (OK, the computers were physically huge, but they were logically tiny). And we had to code everything. Now we can move around preposterous amounts of stuff with a few lines, and we generally use components to sub-contract out lots of detail. Plus, our source code resides thousands of kilometers away from the computers where it runs. (And, by the way, we might well not even know where it executes.) Who knows what we’ll do in the quantum future.
The point is, even if I had the source code from 1972, it would be gibberish to me because it did something totally different than code I write today to accomplish the same thing. Stuff works different.
In short, there is good reason to wonder if future historians will be able to even understand what this archive is about.
qua githubAnd then there is the whole thing about github. Yes, github is extremely popular. This is a huge sample of software. But is it a representative sample? Of the millions of projects included, all but a few are basically trivial.
And, of course, most software is not open source, let along stored in github. The entire idea of “open source” is controversial, contested, and ever changing. This is a sample, of software, but what does it represent?
And then there is the “cultural” context. The archive includes a bunch of “how to” books which you have to wonder about. Telling people 1,000 years from now how computers worked a few years ago seems like a pretty pointless thing. If they have computing, it will work differently than our stuff. If they don’t, then they probably won’t understand, and arguably don’t need to understand our stuff.
And no, the details of say, Python or Ruby, are not timeless gems for the ages. Most of the details are messy and, frankly, it is embarrassing to think they might be examined by future archeologists.
And then there is the “cultural context”. Wikipedia is awesome, but it isn’t everything. And it’s pretty clear that you couldn’t reconstruct our civilization from Wikipedia. It’s an interesting question what you actually would understand, if you just had Wikipedia (even setting aside technical details of language and encoding).
There is also an eclectic collection of other materials, including a tiny handful of fiction. It’s a weird collection, for sure. It like like a 16-year old’s summer reading list. In the UK. In 1999.
This list is supposed to convey “the history and culture of our time”. Hmm.
Why is “I, Claudius” in this list? Even more disturbing, why is “1984” here? Just what would you think about “our time” from “Crime and Punishment”, “The Complete Works of William Shakespeare”, and Homer’s poems translated to English?
There are no git repositories of source code in any of these works. It’s going to take a lot of head scratching to figure out what these have to do with the stuff in this vault. (Because they have nothing to do with it.)
The point is, to accomplish this 1,000 year mission, there are huge, huge semantic gulfs to overcome. Even if the material is technically accessible, I have to wonder what, if anything, future accessors might actually understand.
And here’s a real puzzler for future readers. Github is really, really ethnocentric. It is dominated by the English language, it is part of a sub-culture that shares an array of conceptual ideas about technology, society, and human life.
For example, even if you know a bit about the early twenty first century, you don’t necessarily know anything about, say, how “Linux” got its name. So it’s going to be a hell of a job trying to grok these QR codes, that’s for sure.
Overall, the odds of success are low. And it’s not even clear what this might accomplish if it does “succeed”.
Personally, I consider it a waste of time and resources. (The seed banks are not a wasted effort, though they are just as much a long shot.)
- Jon Evans, “If you don’t make it beautiful, it’s for sure doomed”: putting the Vault in GitHub’s Arctic Code Vault, in Github Blog, September 20, 2022. https://github.blog/2022-09-20-if-you-dont-make-it-beautiful-its-for-sure-doomed-putting-the-vault-in-githubs-arctic-code-vault/
- Liam Tung, Open-source software that lasts a thousand years? GitHub adds to its frozen Arctic Code Vault, in ZDNet, September 21, 2022. https://www.zdnet.com/article/open-source-software-that-lasts-a-thousand-years-github-adds-to-its-frozen-arctic-code-vault/