When the Web Loses Its Memory
As sites like Tumblr perform mass purges, archivists are fighting a losing battle to save disappearing parts of the internet
Diana poured her passionate fandom into Thorvaenn, her Tumblr account, for years. Marvel and Thor in particular became the primary focus for the GIFs she made, for her meticulous recommendations, and for the fanfiction she created involving Thor and Loki — a coupling known as “Thorki” — in various romantic situations for her 20,000-plus followers.
“My blog was super important to me and to my life,” says Diana, who asked to be identified by just her first name. “It was essentially there for me every waking moment. Any time I was bored or restless or wanted to take my mind off things, I could read something or go to one of my friends and chat about it. For as long as I can remember, I was that person, the one who was into something, who couldn’t just watch or read a thing and move on, but who had to be all deep into it.”
“I had that blog for seven years. I poured so much time into it, so much creative energy.”
Then, one day, it was all gone. On November 22, 2018, her blog was deleted without warning. Hundreds of hours of work and craft were gone in an instant. Tumblr erased the blog just before the company announced its ban on adult content, a move that affected huge parts of the site, because, well, Tumblr was home to a lot of kink. Diana maintains she didn’t post the kind of adult content the company was set on banning — defined by Tumblr as “adult content real-life human genitals or female-presenting nipples, and any content [depicting] sex acts” — but Tumblr never replied to her appeal. “I had that blog for seven years,” she says. “I poured so much time into it, so much creative energy.”
As a heavily invested member of the Harry Potter sections of LiveJournal, Eve Elizabeth Moriarty had a different obsession. She clearly remembers reading a piece of fanfiction when she was about 15, shortly after Harry Potter and the Order of the Phoenix came out in 2004, that took place at Sirius Black’s home after he died.
“It might have been the first fanfic I ever read,” says Moriarty, who’s now an academic researching fan studies at Swansea University. “I was sitting at my family desktop computer, which only had dial-up at the time, reading this fanfiction and for the first time realizing the extent of human misery. I went upstairs and cried for hours.”
But now? She can’t find it anywhere, and the same story goes for huge parts of that period of her online life. “That’s a seminal piece of my childhood I just can’t access,” Moriarty says. “Obviously, doing what I do now, it’s an academic concern, too — but for me personally, there’s a real sense of personal, emotional loss.”
Eve and Diana’s stories both illustrate how everything we make and consume online may be just one update away from vanishing forever. We’re not new to the idea of the ephemeral internet, of evaporating Snapchats or deleted social media posts. In fact, the recent push by social media platforms like Facebook into private messaging indicates consumers’ interest in not permitting everything to be retained by the public web. But mass deletions of entire volumes of internet history are something different — and far more significant.
The Tumblr “purge” is one recent example, but it’s not just Tumblr. After Yahoo sold Flickr, its new owner, SmugMug, decided to limit free accounts to 1,000 photos. Anyone who didn’t buy a pro account as of February 5 had their photos deleted. Then there’s the threat hanging over the Neopets site, swathes of which rely on Flash and will be killed off by Adobe by 2020. From the sudden loss of everything uploaded to Myspace before 2016 to the sad little Photobucket icons littering sections of the early 2000s web, what might have seemed eternal when it was uploaded has proven to have a surprisingly brief life.
What do we lose when huge parts of what used to be central to our online experience are wiped out? Embarrassing Myspace photos aside, we lose crucial historical context to how we lived our lives online — which is why a number of institutions and groups have arisen to try to archive the web. Some are professionals; others are volunteers. But what they all have in common is a concern for the historical gaps these shuttered sites leave behind.
The world’s most comprehensive effort is the Wayback Machine, run by the Internet Archive, a not-for-profit in San Francisco. The organization has been saving snapshots of the web since 2001. There’s also the volunteer-run Archive Team, which describes itself as “a loose collective of rogue archivists, programmers, writers, and loudmouths dedicated to saving our digital heritage.” The latter sprang into action when the Tumblr purge began, with volunteers attempting to copy and preserve Tumblr’s NSFW blogs in just 14 days. Each volunteer installs a program on their computer that makes their computer part of a scraping network, archiving content onto the team’s servers. With an estimated 700,000 blogs at risk, it was a mammoth task, and they only managed to capture a portion of the total content in the short time they had.
In some countries, archiving pockets of online culture is done by arts and academic organizations. In the U.K., this includes the British Library, which dates its origins to 1753. Jane Winters is a professor of digital humanities at the University of London, and a big part of her work is the annual snapshot of U.K. web domains that the British Library takes for its archive. “It’s got things like official government publications, individual blogs — a collision of different kinds of content all in dialogue with each other,” Winters says.
Once popularity and profit — or the future chance of profit — begin to wane, there’s just less incentive to sustain a service that’s shedding losers.
Essentially everything marked with the U.K.’s co.uk domain is archived by the British Library, although that means many pieces of content, like videos made by British YouTubers, fall outside the library’s remit because they are hosted under international domains such as .com. Libraries in other countries do similar work. The National Library of France, for example, has archived tweets sent out at the time of terror attacks.
Social media sites in general, however, pose challenges to archivists. “Facebook is very difficult to archive and, I think, makes it very difficult to archive,” Winters says. That’s because every person’s Facebook feed is personalized and different, so there’s no definitive version to archive. “Lots of social media content is falling outside of this archiving activity.”
The companies that create and run these platforms seem to care little about the cultural significance of the content their users create. Sites get sold — like Tumblr to Yahoo in 2013 — or become less popular. Once popularity and profit — or the future chance of profit — begin to wane, there’s just less incentive to sustain a service that’s shedding users.
Some of Tumblr’s long-term users say the site was fading away before the porn ban, and even Diana admits she hadn’t been posting as much when the site was shuttered. “When I realized, after all that crap on Tumblr went down and I saw that this blog went down with it, I thought, ‘Thank god this didn’t happen a year ago, because a year ago it would have killed me.’”
Even so, she felt like her blog was a resource, a starting point for other people who were into that little subsection of Marvel fandom. “I consider fandom history important, and without sounding vain, my blog was a good point for many beginners to go to,” Diana says. “So, even if I wasn’t planning on posting, I wanted it restored.”
Diana has another blog now and mostly posts about the occult drama Supernatural — another show with a huge fan base on Tumblr. But it’s not the same, because she can’t help thinking: What’s the point in being emotionally invested in something if it could disappear tomorrow?