The Internet Archive provides a service called Wayback Machine which lets you view the history of a website. This is very useful when, for whatever reason, a website drops off the Internet.
I used to link to one such site, and when it went away, I changed to pointing at it in the archive. I thought this would be safe.
I was wrong.
The domain has since been bought out by a company that provides those nasty wastes of space that are websites with no content save for a single page covered in contextual adverts. They have a robots.txt.
The Internet Archive, in their "wisdom", have decided to retroactively respect robots.txt files. So all the content provided by the former owner of the domain and stored in the Internet Archive has now been wiped out.
I've deleted the link now, since it pointed to the archive's "Not Found" page, but in future I'm going to use local mirrors of content and not depend on the IA.