I spent most of Saturday afternoon doing something that sounds boring even to me: going through old techpad posts and clicking every single outbound link to see what's still alive. Coffee, a spreadsheet, and fourteen years of my own writing staring back at me. Not exactly a thrilling way to spend a weekend before Christmas, but the guilt had been building for a while.
Here's the thing that got me started. I went looking for a post from early 2012 where I'd written about Instagram getting bought by Facebook, mostly because I wanted to link it from something else I was drafting, and half the links in that post just... went nowhere. A screenshot host that doesn't exist anymore. A commenter's personal site that's now a parked domain selling itself for $12.99/year. A link to a forum thread that 404s. It wasn't even surprising, but seeing it in one place made me actually sit down and count.
Out of roughly 640 published posts, I ran a link checker (a dumb little Python script, requests plus a status code check, nothing fancy) against everything and got back 1,340-something dead links. Some of those are false positives — sites that just block scripted requests and return a 403 even though they're fine in a browser — but I'd guess genuinely dead is somewhere around 800. That's not a typo. Eight hundred links, gone, out of a blog that's only ever had one author.
The usual suspects
A few patterns showed up immediately once I started sorting by year:
- Anything linking to Posterous is dead, obviously, since that got shut down in 2013.
- A huge chunk of 2012-2014 posts link to Google Reader threads or shared items, which stopped existing the same year.
- Twitpic links are dead. Twitpic itself has been dead since 2014, resurrected briefly, dead again.
- A surprising number of 2016-2017 posts link to Vine videos, which is its own small tragedy honestly, some of those clips were genuinely funny and now they just don't exist anywhere I can find.
- Personal blogs from people who commented on my stuff a decade ago. Most of those are just gone, no redirect, no archive, nothing.
The one that annoyed me most wasn't even a dead link, it was the Wayback Machine itself being flaky. I know the Internet Archive got hit with that big breach back in October 2024 and has been rebuilding trust and infrastructure since, and I get it, they're a nonprofit running an insane amount of the internet's memory on a shoestring. But trying to pull archived versions of 200+ dead pages this weekend, I hit timeouts and 503s constantly. Had to run the whole batch twice, spaced a day apart, just to get through it. I'm not really complaining about the Archive as an institution, I think what they do is close to a public good at this point, I'm just noting that depending on it as a single point of failure for the entire internet's history is a little unsettling when you actually lean on it for hundreds of requests in a row.
What I actually did about it
I didn't fix all 800. I'm one person with a day job and this is a hobby blog, not an archival project. What I did was:
- For anything with a Wayback snapshot, swapped the live link for an archive.org permalink.
- For anything without a snapshot, I just struck through the link text instead of pretending it still points somewhere real. Feels more honest than a link that goes to a domain squatter now.
- Left a handful completely alone because fixing them would've meant rewriting the sentence around them, and some 2012-era jokes just don't need that kind of archaeology.
I got through maybe 300 of the 800 before I ran out of patience, which is fine. I'll pick it back up in January probably, or I won't, and either way the blog keeps existing.
What strikes me most, doing this kind of pass every couple of years, is how much faster things rot now than they used to feel like they would. A link from 2012 dying by 2016 seemed reasonable, the web still felt new and companies were still failing fast back then. But I've got dead links now from posts I wrote in 2023. Two years. A service I linked to two years ago is already gone, no redirect, no notice, just silence. That's the part that actually bugs me, not the old stuff, the recent stuff dying this fast.
Anyway. If you've got an old blog and you've never run a link checker against it, I'd recommend setting aside an afternoon before you're much older than I already am.