| The History of Open Source is tied to three | | | | In the early 1980s, the community at MIT AI Lab |
| operating systems: Unix, GNU, and Linux. These were | | | | began to collapse, due in part to computer |
| projects that initially shaped the identity of the open | | | | architecture advances that rendered ITS obsolete. |
| source community, beginning in the 1960s and | | | | Computers that were replacing MIT’s |
| continuing to the present day, and proved that open | | | | PDP-10s had their own operating systems, but none |
| source is a viable software development model. No | | | | were open source. Even getting an executable copy |
| study of open source is complete without | | | | meant signing a non-disclosure agreement. |
| understanding the history of these systems. | | | | At the same time that ITS became obsolete, the AI |
| Unix | | | | Lab community also disbanded. One of the first |
| Multics | | | | people to move away was Brian Reed of Carnegie |
| In the late 1960s, Bell Labs, owned by AT&T, began | | | | Mellon University. Instead of sharing his |
| working collaboratively with General Electric and the | | | | text-formatting program Scribe, with the AI Lab |
| Massachusetts Institute of Technology to create a | | | | community, he sold it to a commercial company, |
| new operating system. The system, known as | | | | which altered the code to insure profits rather than |
| Multics, was to be used in-house at Bell Labs. | | | | communal improvements. Soon, spin-off companies |
| Although Multics was a significant achievement in the | | | | began breaking away from the community. |
| realm of computer science, it was also | | | | Eventually, nearly all of the programmers were hired |
| time-consuming and expensive. Its goals were too | | | | away to work on commercial software projects. |
| lofty for Bell Labs to achieve, and in 1968 Bell began | | | | Richard Stallman was left with a choice: |
| to withdraw from the project (Hauben 1994). | | | | "One: join the proprietary software world, sign the |
| Some of the last people to work on Multics, at Bell | | | | non disclosure agreements and promise not to help |
| Labs, were Ken Thompson, Dennis Ritchie, and Joe | | | | his fellow hackers. Two: leave the computer field |
| Ossanna. For these three and others, the loss of | | | | altogether. Or three, look for a way that a |
| Multics was a disaster. At the time, other operating | | | | programmer could do something for the good. He |
| systems lacked the flexibility and simplicity that | | | | asked himself, was there a program or programs he |
| Multics had promised. Thompson and the rest decided | | | | could write, so as to make a community possible |
| to build a new operating system to suit their | | | | again?" (Rasch 2000). |
| programming needs. After Bell Labs rejected their | | | | The GNU Project |
| requests for a new computer, the group found an | | | | Stallman’s ideals of software were lofty: |
| obsolete PDP-7 computer on which to begin their | | | | he wanted free software for the masses. According |
| efforts. They called the new operating system Unix | | | | to Stallman, the definition of a truly free software is |
| to distinguish it from Multics and began work. | | | | a program that allows users the right to run the |
| Multics was distributed in 1975 to 2000 by Groupe Bull | | | | program for any purpose, modify the program to suit |
| in Europe, and in the U.S. by Bull HN Information | | | | their needs, redistribute copies with open source, and |
| Systems Inc. In 2006 Bull SAS open sourced Multics | | | | distribute modified versions of the program with |
| versions MR10.2, MR11.0, MR12.0, MR12.1, MR12.2, | | | | open source (Stallman 2002). |
| MR12.3, MR12.4 & MR12.5. [1] | | | | Stallman decided to start creating open software by |
| Unix | | | | developing an operating system, the most crucial |
| Work progressed smoothly on Unix throughout the | | | | software for using a computer. He anticipated a |
| early 1970s. The group acquired new computers, | | | | "community of cooperating hackers" that would |
| developed the high-level programming language C, | | | | develop around the project much as in the MIT AI |
| making Unix portable, and created many new tools to | | | | Lab (Stallman 2002). He chose to make the operating |
| make Unix more useful. Within Bell Labs, other | | | | system Unix compatible because that was the |
| departments began to use Unix for a variety of | | | | dominant system at the time. Stallman picked the |
| tasks. Unix eventually became the standard for | | | | acronym GNU for his project, according to a hacker |
| Bell’s computing needs, and a | | | | custom of creating recursive acronyms. GNU stands |
| development support group, called Unix Support | | | | for "GNU’s Not Unix" (Stallman 2002). |
| Group, was formed to provide support for a | | | | Stallman resigned from MIT in January 1984 so that |
| standard version of Unix. | | | | MIT would have no claim on distributing GNU. He |
| Even in its infancy, word of Unix was spreading | | | | would avoid GNU becoming proprietary software at |
| throughout the computing world. Unix was particularly | | | | all costs. However, he was invited by the head of |
| appealing to the academic computer science | | | | the MIT AI Lab to continue using the MIT facilities. |
| community. Academic institutions were able to | | | | In 1985, Stallman founded the Free Software |
| purchase licenses for the Unix source code very | | | | Foundation (FSF), a tax-exempt charitable |
| cheaply. Government and commercial licenses were | | | | organization, to support the free software |
| much more expensive. | | | | development being done on the GNU Project. |
| In November of 1973, Professor Bob Fabry of the | | | | Stallman also contributed to the project by writing a |
| University of California at Berkeley attended a | | | | multiple language compiler known as GCC, a debugger |
| presentation on Unix at Purdue University. His interest | | | | (GDB), a text editor (GNU Emacs), and other |
| was piqued, and Fabry convinced Berkeley to | | | | software. |
| purchase a PDP-11/45 computer capable of running | | | | In order to ensure that GNU would remain open |
| the current version 4 of Unix. In January of the next | | | | source in future, Stallman created the GNU General |
| year, Version 4 Unix was installed with the help of | | | | Public License (GPL). The GPL specified "that users of |
| Ken Thompson of Bell Labs. | | | | the source code could view, change, or add to the |
| In the fall of 1975, Ken Thompson decided to take a | | | | code, provided they made their changes available |
| one-year sabbatical from Bell Labs to teach at | | | | under the same license as the original code" (GNU |
| Berkeley, his alma mater. The Computer Science | | | | General Public License). Stallman received the |
| department had just purchased the new PDP-11/70 | | | | MacArthur fellowship, which entails a stipend of |
| computer, and Thompson helped to install the latest | | | | $500,000, in 1990 for his work with GNU, the GPL, |
| version of Unix, Version 6, on it. Two graduate | | | | and the FSF. |
| students, Bill Joy and Chuck Haley, also arrived in | | | | The GNU operating system continued to grow |
| 1975, and began working on a Pascal compiler written | | | | throughout the 1990s, developing piece by piece. |
| by Thompson, a new text editor, and improvements | | | | Each piece was implemented on a Unix system, so |
| to underlying parts of the Unix system itself. | | | | that components could be completed and distributed |
| Other programmers began to take interest in the | | | | before the entire system was released. By 1990, the |
| new Pascal compiler at Berkeley, and during the year | | | | only major piece missing from the system was the |
| of 1977, Joy began to distribute the "Berkeley | | | | kernel. A kernel is the core of an operating system |
| Software Distribution," an open source version of | | | | "that provides basic services for all other parts of the |
| Unix containing the improvements and additions made | | | | operating system" ("Kernel"). |
| at Berkeley. BSD was sold for a nominal fee to | | | | Stallman’s team began work on a kernel in |
| people who had already obtained a Unix license from | | | | 1990, called Hurd. However, work is progressing |
| AT&T. In mid-1978, Joy put together the "Second | | | | slowly. |
| Berkeley Software Distribution," or 2BSD, which was | | | | Linux |
| distributed the following year. Distribution jumped | | | | Minix |
| from about thirty copies with the original BSD to | | | | In 1987, a professor, Andrew Tanenbaum, invented |
| about seventy-five copies with 2BSD. | | | | Minix, a clone of the Unix operating system to be |
| Improvements continued, increasing portability, | | | | used for educational purposes. Minix was not the |
| improving memory usage, and implementing new | | | | most sophisticated of operating systems, but its |
| tools. In December 1979, 3BSD was released, and | | | | appeal to programmers worldwide was that all 12,000 |
| nearly 100 copies were shipped. At this time, with the | | | | lines of C and assembly were available to be studied |
| breakup of Bell, the price of Unix licenses for the | | | | and tinkered with (Hasan 1999). |
| academic community began to increase. AT&T | | | | Linux |
| shifted management of Unix to a new group and | | | | In August of 1991, Linus Torvalds, a 21 year old |
| began emphasizing proprietary versions of Unix | | | | Computer Science student at the University of |
| (Dibona et al.: McKusick 1999). The first of these new | | | | Helsinki, posted to the Minix users newsgroup that he |
| releases was System III in 1982, followed in 1983 by | | | | was working on a new, free operating system, |
| System V. Berkeley moved to fill the void Bell had | | | | adding parenthetically that it was "just a hobby, |
| left in open distributions and continued to release | | | | won’t be big and professional like GNU" |
| further versions of BSD, using a new open source | | | | (Newitz). What Torvalds was actually creating was a |
| software license known as the BSD License. | | | | kernel, the core of an operating system. |
| AT&T did not begin heavy commercial promotion of | | | | In 1992, the completed Linux kernel was combined |
| Unix until the mid-1980s. What they found when they | | | | with the incomplete GNU operating system, resulting |
| ventured into the market was that many vendors | | | | in a working open source operating system. |
| were already selling their own proprietary version of | | | | According to Stallman, "It is due to Linux that we can |
| Unix. The issue of the day was which version of Unix | | | | actually run a version of the GNU system today" |
| would become dominant. In 1987, in an effort to | | | | (Stallman 2002). In later years, this combination of |
| unify the market, AT&T formed an alliance with Sun | | | | GNU and Linux along with other free software |
| Microsystems, a strong supporter of BSD. In | | | | exploded in popularity and became commonly known |
| response to the move, several vendors created the | | | | simply as Linux. |
| Open Software Foundation (OSF) to support their | | | | In September, version 0.01 of the Linux kernel was |
| own open source versions of Unix. AT&T and Sun in | | | | released on the net, and enthusiasm began to rise |
| turn formed Unix International. Thus the "Unix Wars" | | | | around the project (Hasan 1999). On October 5th, |
| began (The Open Group 2001). | | | | Torvalds sent a formal call for volunteers to the Minix |
| The Unix Wars | | | | newsgroup, saying "Are you without a nice project |
| Throughout the remainder of the 1980s and into the | | | | and just dying to cut your teeth on a OS you can |
| 1990s, the Unix Wars raged. During this time, many | | | | try to modify for your needs?" (Newitz 2000). |
| different versions of Unix were released, both | | | | Torvalds’ plea appealed to many |
| proprietary and open source. In 1991, AT&T spun off | | | | programmers’ senses of curiosity and |
| Unix System Laboratories, which passed through | | | | excitement. By December, version 0.10 was released, |
| several hands before being bought in 1995 by Santa | | | | still as a bare-bones kernel. |
| Cruz Operation (SCO). | | | | And soon after version 0.12 was released, Linux was |
| Three distinct versions of BSD emerged from the | | | | then licensed under the GNU General Public License to |
| Unix Wars: FreeBSD, known for its simplicity, stability, | | | | ensure that the source would be free to all (in fact |
| and ease of use; NetBSD, known for its portability | | | | the readme for that release stated that it will soon |
| and research-oriented environment; and OpenBSD, | | | | be licensed under that license. Linus was just checking |
| known for its high level of security and stability | | | | for licensing issues before doing so). In the ensuing |
| (Howard). Other versions of open source Unix are | | | | years, thousands of people began working with |
| available as well, including a version from SCO itself | | | | Linux, helping to improve the kernel itself or writing |
| and Darwin, the foundation on which | | | | software for use on Linux systems. |
| Apple’s Mac OS X is built. | | | | Throughout the 1990s, as Linux swelled in popularity |
| The GNU Project | | | | and became more and more sophisticated, vendors |
| MIT | | | | began distributing it commercially. Although Linux was |
| In 1971, Richard Stallman, a Harvard undergraduate | | | | and is free and open source, vendors such as Red |
| student, began working at the MIT Artificial | | | | Hat, Novell and Mandriva have gathered it into a |
| Intelligence Lab, primarily on the ITS, an operating | | | | format that more closely resembles other |
| system unique to the computers at MIT. The | | | | contemporary operating systems. With the addition |
| community at the MIT AI Lab was a small group of | | | | of graphical user interfaces and other user-friendly |
| programmers who improved code by passing it back | | | | features, these distributors were able to profit from |
| and forth; in other words, the group’s | | | | selling open source bundled into a product that |
| software development basis was open source | | | | everyday users wanted and could use with ease |
| (Rasch 2000). | | | | (Newitz 2000). |