It’s rather pleasing to see that ‘link rot’ and dead connections have been part of the internet ever since the start. As shown in RFC-315 reporting on a week’s worth of server status reports in February 1972, as read by Darius Kazemi. Of course the underlying notion of the internet as distributed system is that if a node fails, the rest can continue to work. This was fundamental to ARPANET. What’s pleasing to me is the fact that such robustness was needed from the start, that over half of servers could be ‘dead’ and the network still be seen to exist. That intermittence was not just a theoretical requirement but an every day aspect in practice. Intermittence has been on my mind a lot in the past few weeks.