Link Rot: A Problem for the Supreme Court and the Web
Link rot is the phenomenon where links on a website lead to websites and other web resources that have become unavailable. This is due to reasons like the domain name changing, domains not being renewed, purposeful removal of web content, etc. All of this creates a situation where search engines sometimes give useless results that may lead to a dreaded “Error: 404 page”. If we think back to the bow-tie model of the web, the in component of the model is still the same as before but the out component of the model has been significantly reduced because of link rot. In other words, the number of nodes leading out of a website goes down significantly.
For a more specific example of how detrimental link rot can be to search engine optimization, we can look at a 2014 Harvard study. The study determined that about half of the links in US Supreme Court opinions are dead links. As the Networks textbook explains, citations are essential for these decisions because they explain and give precedents. With dead links, it’s harder for politically-minded citizens, legal scholars, and other people to understand a Supreme Court opinion. If a set of opinions with high authority scores align with well with the judgements of highest quality, then this asks the question of how reliable these scores actually are. If the hubs and authorities scores were calculated at a time when most of the link rot hadn’t happened yet, then the scores are reliable. (The average lifetime of a web page is 100 days. However, it would be reasonable to assume that web pages detailing Supreme Court opinions would be longer. On average, the court hears 80 cases a year and decisions of this magnitude probably don’t get amended often.) However, if the link rot had already happened, then the scores might not be reliable.
Suppose there is an authority A and hub H. Hub H links to authority A but that link is a dead link. Another hub I trusts hub H and wants to link to authority A, but it only finds a dead link. So this dead link deters other hubs from linking to authority A which should lower its authority score. The dead link to authority A isn’t because of a direct lack of effort from authority A, but dead links can certainly affect hubs and authorities scores. It could be argued that these scores would be more accurate if there weren’t dead links which would be beneficial for the end-users.
links:
http://www.newyorker.com/magazine/2015/01/26/cobweb