This article comes from the 36-year-old compiling team, the translation team, the translator boxi, and the love fan is authorized to publish.

Hypertext links are the foundation of the World Wide Web. But without the 404 error, the Internet we are familiar with will not have the scale and prosperity of today. Jesse Dunietz reviewed the birth history of the 404 error and gave it a philosophical high rating: 404 on the World Wide Web, like 0 to math. Useless, it is for big use. Originally published in “Popular Machinery” titled: How the 404 Error Created the World Wide Web

404 errors are obviously innovations, but without it, the Internet you know and love can’t.

It is the scourge of every surfer on the Internet. It is the scary blackboard on the Internet that you are afraid to hear. As long as you click on any link whose creation date goes back to before 2005, click on almost any link that dates back to 2005, and the following line appears almost noCan be avoided: “HTTP 404 error: page not found.”

As long as people who have been online, you should be familiar with 404 errors. This is how the web server tells you that it doesn’t work. But little known is that the WWW exists because of this mistake.

History of hyperlinks

Let’s talk about hyperlinks first. We tend to think that what we call a web – a collection of a large number of linked web pages – is the product of an Internet connection. Or in other words: First there was a communication network that allowed computers to exchange data, and then we built an interconnected labyrinth filled with documents and videos of A-Cats and dogs. In fact, the other way around is correct. Hypertext or the idea of ​​having a track to link to text in other content is decades before the idea of ​​a networked computer.

▲Vannevar Bush’s conceptual map of the microfilm-based memex information machine, from his 1945 article in the Atlantic Monthly.

Hypertext dates back at least to 1945, when technology pioneer Vannevar Bush proposed a hypertext-enhanced microfilm machine called “Memex (Extension of Memory).” Bush envisions leaving a small piece on the edge of the microfilm, and then, according to the user’s instructions, memex can place the address code of the relevant film on the edge of the film. At any time thereafter, anyone viewing the same microfilm can immediately open the linked panel.

But Bush’s thoughts were too advanced, and until the 1960s, his ideas were always a fantasy. With the advent of digital computers, true hypertext quickly became a reality. IT legend Ted Nelson drew on Bush’s ideas and conceived an ambitious hypertext concept called the Xanadu project, although it was not until 1998 that it was partially implemented. However, by the late 1960s, Nelson had worked with people to develop a less complex hypertext system that would support links within documents.

At the same time, Douglas Engelbart, one of the early leaders in the field of human-computer interaction, is studying its revolutionary NLS (oNLine System). NLS has many groundbreaking features, one of which is that the system allows users to jump through the document using hyperlinks. So, as early as the mid-1980s, hypertext systems appeared in Nelson, Engelbart, and their subsequent work.

The modern network is slowly forming

However, these systems have certain limitations. The biggest limitation is that they are only valid on a single computer. For example, Apple’s HyperCard maintains a scratch card database, but the database can only be linked to other sticky notes cards on the same device. But with the rise of computer networks, it is natural to extend the link from a document on one computer to a document on another. Even so, until 1989, CERN contractor Tim Berners-Lee invented the World Wide Web.

In 2009, Berners-Lee said in a TED speech recalling the birth of HTTP: “What is frustrating is that there is too much potential not released. Everyone has a lot of documentation on disk. If you imagine all of this If they belong to a large virtual document system that is set up in the air (such as the Internet), then everyone’s days will be much better.”

However, this idea has to take root in a wide range, and there is still something missing. This thing is a 404 error.

Before Berners Lee, hypertext systems usually ensure that each link points to a place. All new links are added to a central database of recorded documents and links. If the target of the link has been changed or deleted, the database must also update the link accordingly.

If you imagine that all of this belongs to a large virtual file system that is set up in the air (such as the Internet), then everyone’s days will be much better.

Keeping the consistency of hyperlinks is very helpful for users. This is not a trivial task when all data resides on a single computer or a small network. butIf you are on a large computer network, you will need a central authority to register all the documents and links. No database can update all the links in the world in real time.

This problem has hardly received much attention for a while. The focus of most researchers is still on record cards, help applications and other small systems. Some projects do allow one-way links from one machine to another without a central authority, but there is a premise that these links are maintained as part of a team-critical document creation process. .

The results show that there is actually a much simpler way to solve this problem.

The birth of “404 Not Found”

Berners – Lee presents a wonderfully easy way to verify links: no verification required.

In the beautiful new world of the World Wide Web, the only place to store information about a link is the document containing the link. If the target document has been moved or changed, it will have to be updated by the linked document. Or let go, no matter what you do.

Of course, this means that the link to the location may not exist. As a result, a 404 error has occurred. Berners-Lee supports the concept of missing content, and he also gives the official error code that should be returned when an error occurs.

So where is the number of 404 coming from? It sounds like it’s casual, but it’s not the case. Berners – Lee’s Hypertext Transfer Protocol (HTTP) has dozens of status codes. Codes starting with 4 are for client-side errors, and requesting a non-existent address (the “04” portion) is just one of many ways that can be messed up when browsing.

In a sense, 404 is like hypertext and 0 is mathematical: of course, this seems obvious, but formalizing it and building up the concept revolutionizes the rest of the system.

With the innovation of Berners-Lee, hypertext documents can be linked to any other document that they know the address. For example, if you want to link to this article, you don’t need to get my permission or coordinate with Volkswagen Machinery. You can do it. If we modify the URL without setting a redirect, or if we delete the article, you will receive an error message.

This independence is one of the key factors behind the web’s thriving business. Within a few years, the world has fallen into a crazy web page and a one-way link to the ocean.

When a 404 error occurs, the website displaying a custom page is a fashion. Even Hillary Clinton’s campaign site has a 404 comedy action movie.

In a sense, 404 is like hypertextAnd 0 to math: Of course, this seems obvious, but renaming it and building up the concept has revolutionized the rest of the system.

Although this new solution is very good, there are some problems, mainly because the link is invalid. As time goes by, the pages migrate, the site replaces its content, and even the entire site goes offline, causing the chain pointing to them to contact the reef. The study found that at least 50% of published links will expire within five to ten years. But at least some websites have made some creative attempts to treat us with a custom 404 error page.

There are techniques to avoid invalid links, such as carefully choosing a link URL, or a bigger package: archiving everything. But for the Web, occasionally a face 404 caused by the face, but the endless news, knowledge and cat model due to a little price.