A simple idea with a long history
Hypertext means text that contains links to other text, so that a reader can jump between related passages instead of reading in a straight line. The web made this familiar to everyone, but the idea and several working systems existed decades earlier. Printed books already had footnotes, indexes and cross-references, which serve a similar purpose, and the new idea was to make the jump instant and, eventually, to let readers and writers create their own paths.
Vannevar Bush and the memex
A frequently cited starting point is an essay by the American engineer Vannevar Bush, published in 1945 under the title “As We May Think”. Bush worried that scientific knowledge was growing faster than people could use it. He described an imagined desk-sized device, the memex, in which a person could store books, records and notes on microfilm and connect items to one another with “trails” of associations.
The memex was never built as described, and it relied on microfilm, not computers. Its importance lies in the concept: personal, associative paths through information, which could be saved and shared. Later pioneers often credited the essay as an inspiration, although historians note that others were also exploring related thoughts at the time.
Naming and visions in the 1960s
The word hypertext is generally credited to Ted Nelson, who used it in the 1960s. Nelson imagined a worldwide system of connected writing in which documents could quote one another and credit could be tracked. His long-running project, called Xanadu, was described in books and discussed for decades, but it never became a widely used product in the form he envisioned. It nonetheless influenced how many people thought about links, versions and ownership of text.
Around the same time, Douglas Engelbart and his team at the Stanford Research Institute worked on a system for augmenting human thinking, which included linked and structured documents. A public demonstration in 1968 showed screen editing, links and a mouse, and is often called a landmark in the history of interactive computing. It is closely tied to the development of the graphical user interface and pointing devices.
Research systems at universities
In the late 1960s and 1970s, university teams built working hypertext systems on mainframe computers and later on workstations. One project at Brown University, led by Andries van Dam and colleagues, produced editing and reading systems that were used for teaching and documentation. Such systems were typically used by small groups, ran on expensive equipment, and were not connected to a global network.
These projects explored questions that the web later had to answer, including how to show a reader where they are, how to deal with broken links, and how to let people annotate.
Systems for ordinary computers
In the 1980s, personal computers made hypertext practical for more people. Several products appeared. Among the best known was HyperCard, released by Apple in 1987. It organized information into “cards” with buttons that could jump to other cards, and it included a simple scripting language. Teachers, hobbyists and small developers used it to build tutorials, databases, games and reference tools. Because many Macintosh computers included it, it reached a broad audience.
Other products and research prototypes from this period were used for technical manuals, help systems and educational materials. Some, like a widely used help format in operating systems, put hypertext in front of users who may not have realized what it was. CD-ROM encyclopedias and multimedia titles also used linked articles, an early step in the move away from printed encyclopedias.
How these systems differed from the web
Most pre-web hypertext was closed in at least one important way. Documents lived in a single system or file, and links were typically managed by the software, often in both directions. Several features common in research systems were missing from the web’s first design:
- Two-way links. In many earlier systems, a page could show what linked to it. The web’s links point only one way.
- Central control or databases. Earlier designs often used a central store to keep links consistent, which made broken links less likely but made wide, open participation harder.
- Rich authoring. Many systems let readers edit or annotate. Early web browsers focused on reading.
The web’s authors made trade-offs. A one-way link needs no permission and no central record, and a page can be created anywhere. The cost was that links break and pages cannot easily know who points to them, a gap that later search engines partly filled by gathering link data.
The step to the web
When Tim Berners-Lee proposed his information system at CERN, he was aware of earlier hypertext work, and he combined it with existing internet protocols and a simple page format derived from document markup practice. The combination of open standards, free access to the software and the network’s reach helped it spread. Credit for the web is therefore shared with a long line of earlier ideas, with the new part being the way the pieces were joined and made freely usable.
What remains from the earlier work
Many of the earlier ideas have returned in other forms. Wikis let readers edit linked pages, and notes applications often support two-way links. Citation tools track references, and versioning systems keep a history of changes. Some of Nelson’s goals, such as tracking credit, are still debated. Understanding the earlier systems helps explain why the web works the way it does, and what trade-offs it chose.