The text editor Emacs turns 50 in 2026. An article published by MIT Technology Review on October 20 looks back at the history of Emacs, which was born in 1976 at the Massachusetts Institute of Technology's (MIT) Artificial Intelligence Laboratory and has remained in use ever since. However, the GNU Emacs that is widely used today is a different implementation, one that Richard Stallman began developing in September 1984. The same program has not been running for 50 years.
Why, then, has Emacs lasted so long? The answer appears to lie less in any particular source code than in a shared set of key bindings and a development culture in which users improve features themselves and share them. The history of Emacs shows that a piece of software's lifespan does not necessarily match the lifespan of its implementation.
Emacs at 50: Development of the Current GNU Emacs Began in 1984
The origins of Emacs go back to TECO, a text editor Dan Murphy developed for the PDP-1 in 1962. TECO edited text by combining single-character commands, and in 1971 Carl Mikkelsen added the Control-R feature, which immediately reflected editing results on screen. Stallman, influenced by "E," a full-screen editor developed at Stanford University's AI laboratory, then improved TECO and in 1976 developed it into a structure in which a macro could be assigned to each key.
This mechanism let MIT researchers freely add commands, but it also created a problem: every user operated the editor differently. Guy Steele Jr. took the lead, working with David Moon and Stallman to establish a common command system. Key assignments such as Meta-F, which moves forward one word, and Meta-B, which moves back, were standardized in this process. Steele later recalled that the goal was not to build a superior new editor but to solve a social problem within the lab.
However, the Emacs born in 1976 and today's GNU Emacs are not the same implementation. According to the GNU Project's official history, early Emacs was implemented as a collection of TECO macros. GNU Emacs, by contrast, is a new implementation whose development began in 1984, built on C and Lisp. Stallman himself has said he started work on GNU Emacs in September 1984 and that it was in usable shape by early 1985.
GNU Emacs's official version history records that version 13, the first public release, appeared on March 20, 1985. In version 16.56 on July 15, 1985, code developed by James Gosling was removed for copyright reasons. In other words, GNU Emacs itself underwent major changes to its implementation at an early stage.
As of October 2026, about 42 years have passed since GNU Emacs development began. That is roughly 84% of the 50 years since 1976, but this compares only elapsed development time; it does not mean that 84% of the current code survives from 42 years ago. Moreover, in a 1981 paper, Stallman said Emacs development began in 1974, so the length of the history also changes depending on where one places its starting point. How much pre-1984 code remains in the current source needs separate verification, but at the very least, the 50th anniversary does not represent the period of a single continuous implementation.
Key Bindings and a Culture of "Sharing" Outlived Changes in Implementation
What set Emacs apart from many editors of the time was its extensibility: users themselves could add and change features. In his 1981 paper, Stallman emphasized that users could add new commands and modify existing ones while continuing their editing work. By sharing user-written extensions as libraries and folding the best ones into the base system, the whole of Emacs could be improved continuously. This mechanism can be seen as a forerunner of the open collaborative development widely used in software today.
Sharing in early Emacs, however, came with its own rules. In the 1981 manual, Stallman said Emacs was free to use, but he asked users to join the "EMACS software sharing commune" and send back any improvements they made. In a talk Stallman gave in later years, he likewise explained that he gathered improvements and incorporated them into new versions so the whole community could use them.
Not every implementation bearing the Emacs name followed the same sharing rules. In 1978, Bernard Greenberg developed a separate implementation for Multics, which Honeywell sold commercially. On May 6, 1983, Gosling's implementation became "UniPress Emacs," a commercial product priced at $395. Implementations that shared a common operating scheme but could not be freely redistributed did exist.
This experience also connects to the idea of "copyleft" that Stallman later adopted in the GNU Project. According to the GNU Project's explanation, copyleft is a mechanism that permits use, modification, and redistribution of a program while not allowing new restrictions to be added to published improved versions. The GNU General Public License (GPL) is its best-known license. Unlike in early Emacs, however, there is no obligation to send improvements back to Stallman himself. What carried over was not any specific distribution terms but the idea of sharing the results of improvements with a community.
This development culture also shows in the history of Japanese-language support. On December 15, 1989, "NEmacs 3.2.1," a derivative capable of handling Japanese multibyte characters, appeared. With "MULE 0.9.0b," released on March 4, 1992, support for input methods for Japanese, Chinese, Korean, and other languages advanced, and on August 1, 1993, "MULE 1.0," based on GNU Emacs 18.59, was released. A family tree of Emacs compiled by Jamie Zawinski also records the appearance of these derivatives. Users developed the features they needed on their own while maintaining Emacs's operating scheme.
The inheritance of key bindings has spread to software other than Emacs. On macOS, for example, Emacs-derived operations such as Control-A to move to the beginning of a line, Control-E to move to the end, and Control-F and Control-B to move by character work in many text input environments. In some environments, Control-Option-F and Control-Option-B move by word. Apple ended the standard bundling of GNU Emacs with macOS in 2019, but its official keyboard shortcut list still includes many Emacs-derived operations.
Improvements to Emacs itself continue as well. GNU Emacs 31.1, released on August 24, 2026, enabled mouse operation in supported terminals by default and added a "newcomers-presets" theme that bundles alternative settings for new users. This theme does not wholesale replace previous settings; it is a mechanism users can choose to apply. Emacs is making improvements for new users while preserving its long-standing operating scheme.
What Sustained Emacs for 50 Years: Code, or Design Philosophy?
There may be objections to viewing Emacs's 50 years merely as a history of key bindings and a sharing culture. Stallman played a central role in both the TECO-based Emacs and GNU Emacs, and he carried over the design philosophy that prizes extensibility. If software was rebuilt on the same developer's philosophy, it is not unnatural to count its history as that of a single product even when the implementation changed. Steele also highly praises Stallman's achievement in developing the editor on two different foundations, TECO and Emacs Lisp.
Still, a defining feature of Emacs is that its operating methods and design philosophy spread independently of any developer or implementation. Multiple researchers, including Steele, were involved in standardizing key bindings, and Japanese-language support developed in independent derivatives. The operating scheme has also been absorbed into software other than Emacs itself, as in macOS. Given this history, it seems that what supported Emacs's longevity was not any particular code but a mechanism by which users share operating methods and improve features as needed.
That said, points remain to be verified. If a detailed examination of the current GNU Emacs source code shows that a considerable amount of pre-1984 implementation has been carried over, code continuity would also need to be reassessed as a reason for its longevity. And if standard key bindings such as Meta-F and Meta-B were substantially changed in a future major version, the question of how far the traditional operating scheme is maintained would arise. Future changes, however, would not negate the continuity of the past 50 years itself.
Emacs, born in an MIT lab in 1976, has remained in use for 50 years through multiple implementations and derivatives. Its history shows that software's value lies not only in individual code but also in the operating schemes and development culture shared among users. Even now, as GNU Emacs 31.1 advances support for new operating environments, Emacs keeps changing while preserving its traditions.