The internet is the road, the web is what travels on it
People often use “the internet” and “the web” as if they mean the same thing. They do not. The internet is the global system of connected networks that carry data between computers. The web is one service that runs over it: a collection of linked documents and applications that your browser retrieves using a particular set of rules. Email, online games, file transfers and many phone apps use the internet without being part of the web at all. The distinction also appears in the myths series, in web versus internet.
The web’s history, then, is mainly a history of an idea about information, more than about wires. It is a story about linking things together.
Hypertext before the web
The notion of connecting pieces of text by association, not by order, has a long intellectual history. Several milestones are usually cited.
- Vannevar Bush, 1945. In an essay called “As We May Think,” Bush, an American engineer and science administrator, described an imagined desk-sized device called the Memex, in which a person could store books, notes and records and link them along trails of association. The Memex was never built as described, but the essay influenced later researchers.
- Ted Nelson, 1960s. Nelson coined the word “hypertext” and began a long-running project called Xanadu, which aimed to create a universal network of linked documents with features such as tracking of quotations. Xanadu was ambitious, and a full version never reached ordinary users, but its ideas are still discussed.
- Douglas Engelbart, 1968. At the Stanford Research Institute, Engelbart and his team demonstrated an online system with linked documents, a mouse and collaborative editing in a famous public demonstration. His work is part of the history of the computer mouse.
- Hypertext systems in the 1980s. Academic and commercial products appeared, including Apple’s HyperCard in 1987, which let ordinary Macintosh users make stacks of linked cards. Their weakness was that each ran on one computer or within one closed system, so links could not usefully point elsewhere. Their strengths are covered in the article on hypertext systems.
Berners-Lee knew of this tradition. What he added was not hypertext itself but a way to make it work across a large, open and unmanaged network.
What was new in the web
Three decisions in the web’s design explain its success, according to many historians of the field.
- Simplicity. The format for writing pages, HTML, was small and could be learned quickly, as the story of HTML markup shows. Anyone with a text editor could publish.
- Openness. The core specifications were published, and no single company controlled them. In 1993, CERN stated that the technology could be used freely.
- One-way links that need no permission. A page could link to any other page without asking its owner. Earlier hypertext systems often tried to keep links two-way and consistent, which is difficult at scale. The web accepted broken links as a fair price for growth.
The result was messy but able to grow without central planning. The invention itself, at CERN around 1989 to 1991, is told in the separate piece on the development of the web.
Early spread
Between 1991 and 1993, the web was a small part of the internet. Other systems competed for attention. Gopher, from the University of Minnesota, organized information into menus and was popular among universities for a time. WAIS offered text searching across databases, and FTP archives held software and documents. The web’s advantage was its flexibility, since a page could mix text, links and, eventually, images.
Early pages tended to be plain: lists of links, papers by physicists, university department pages and personal home pages. Web servers multiplied from a handful in 1991 to many thousands during the following few years, though precise figures from that time are uncertain. The introduction of graphical browsers such as Mosaic in 1993 made pages more attractive, and commercial interest followed.
By 1994, the first large web directories and early search engines had appeared to help people find things, and the World Wide Web Consortium was founded to coordinate standards. Companies began to put brochures, then shops, on the web. By the middle of the 1990s, the web had become the most visible part of the internet for ordinary users, so much so that many people started to treat the two words as synonyms.
Limits and trade-offs
The web’s openness has costs. Pages move or vanish, so links break. Anyone can publish, so quality varies widely, and misinformation and scams spread alongside useful material. The original design had no built-in way to confirm who wrote a page or to protect private data, so these features were added later and unevenly. And although the web is open in design, in practice much of it is now concentrated in a small number of large services.
A milestone, not a starting gun
The web did not create the internet or the idea of linked information. It combined an old dream of associative reading with a new, open network, at a time when personal computers were becoming common enough to make that combination useful. That history helps explain why it succeeded, and why its original assumptions about trust and openness have required repair ever since.