Pete's Log: Link survival
Entry #1893, (Meta)(posted when I was 42 years old.)
I fell down another silly rabbit hole that I needed to see through to completion. This one took up both of JB's weekend naps.
I make regular use of the random log entry button to revisit old entries. And an external link on one of those random entries caught my eye. So I clicked it, and sadly it no longer worked. Having recently added the "20 Year Club" to my links page, this made me wonder how external links in log entries have fared the test of time.
I wasn't much in the habit of adding external links early on. While 1998-2002 account for 70% of all log entries, those entries only account for 15% of all external links. In total, there were almost 400 external links over the years, of which 117 no longer appear to work. Those links break down as follows:
| Error | Count | Notes |
|---|---|---|
| Not Found | 39 | e.g. 404 errors |
| Host not found | 30 | Mostly domains are no longer registered, but sometimes just the host no longer exists, e.g. www.lsc.nd.edu |
| Bad Redirect | 15 | URL redirects to root of site or other site unrelated to the URL |
| Misc Error | 12 | e.g. blank pages or 5xx errors |
| Spam | 9 | Including "domain for sale" sites |
| Timeout | 6 | |
| Not Authorized | 4 | e.g. 403 errors |
| Defunct | 2 | Former owner has left a goodbye message |
By decade, the link survivorship rate is
- 1990s: 11% (1/9)
- 2000s: 58% (124/213)
- 2010s: 87% (119/137)
- 2020s: 100% (23/23)
I'm debating going back and editing links to point to archive.org where available. Because link rot makes me sad.
A few other observations: my first HTTPS link wasn't until 2015. I found this surprising.
My first link to Wikipedia was in 2005. Since then I've linked to Wikipedia 91 times (3 of which were to the German Wikipedia). So Wikipedia accounts for almost a quarter of all links. Sadly, even Wikipedia wasn't fully immune to link rot. I had one link (Hammerschlagen) that was deleted and another (Eisbach) that was moved and replaced with a disambiguation page. So should I go back and update my link to point to the correct page instead of the disambiguation page? These things keep me up at night.
My first link to YouTube was in 2006, with the now-quaint note "I know everybody knows youtube by now, but I want to point out how much I love it." Ah, those were simpler times. I linked to YouTube 19 times, of which 7 links no longer work, which I found to be a surprisingly high casualty rate. And that's not even counting the fact that several of those links were actually Flash embeds that don't work anymore. I guess if I'm going to do some cleanup, I should replace all those embeds with working links.
There were also two links to Geocities. Oh Geocities.