The Internet is(n't) forever
I’m in my 30s, as of the time of writing this post. I was a part of the Homestar Runner and Newgrounds generation, cutting my teeth on a very weird web drunk on its initial forays into dynamic websites.
The internet wasn’t just a few sites linking between each other, web rings were sometimes way more useful for finding things than search engines, and making your own site on angelfire or geocities was how you supported your fandom. I remember the start of social media and the warnings that “the internet is forever”, that what you put online is going to last forever so you better be careful.
That’s not really what’s happened, is it? I mean, yes, you can go to someone’s social media account and trawl it for something embarassing. That prediction came true. But, the general sentiment, that everything online is going to be there and visible forever hasn’t. I was a big tumblr user back in the early ’10s. Trying to actually find viral posts from tumblr that had literally millions of replies and reblogs more than a couple of years later is nearly impossible. You can search for phrases that exactly appeared in the post and still not find it. It’s bizare. Even trying to find the context of what people were responding to in old blog posts is far more difficult than you’d expect unless there is an explicit link included. Searching or perusing doesn’t get you the context you need. Too much of it is carried in very ephemeral social media context, a kind of internet version of short-term memory.
Short-term internet memory is when we recognize user names, blog names, sites. It’s when we all remember viral posts and memes in ways where no one has to name the context anymore. The really big ones are likely to be preserved. I don’t think we’re in danger of entirely losing the context of “I’ll face God and walk backwards into Hell” as a meme. But what about smaller bits of context, things more community specific? That’s what worries me. I worry about the passage from short-term internet memory, where people are doing a lot of the work of carrying context, and long-term internet memory where the structure of the documents carries the context within it. I’m using short-term and long-term memory here as an analogy to our own memory. One of the big problems with our long-term memory is that, on some level, we don’t “lose” memories. It all stays in there. The problem is one of indexing. To forget something isn’t to have the information lost from our body, it’s to have the ability to access it lost. I think that’s very similar to what’s happening here with the internet: the internet is always there but never rememberable.
This largely means links, although links themselves are still vulnerable to breakage over time. I’m ignoring the issue of link breakage to a certain extent because there are resources like the Internet Archive that are at least attempting to deal with that problem.
I’ve noticed over the past few years that search engines in general are getting less useful. They’re prioritizing results in ways that are geared towards promoting commerce or recentness over accuracy. They’re still useful short-term memory aids, helping us find things that have happened in the past year or two, but I don’t think they’re useful for reconstructing long-term memory.
I think one of the problems is that much of our communication online is on social media. Social media is such a crowded, fast moving, space that it’s practically redundant to explain the context of events. Moreover, it’s frequently not even safe to link to things that are happening lest you pick up the wrong kind of attention from bad actors. Instead, people frequently use screenshots which—of course—just add to the difficulty of keeping context alive.
What would a robust, searchable, long-term internet memory look like? I think it would have to be something almost like the internet archive but, perhaps, rather than archiving websites with a timeline per site it’s more like creating explorable time slices of pruned parts of the web. How would this be generated or maintained? I have no idea yet!