History of Unix

from Wikipedia, the free encyclopedia
Rough overview of the development and ancestry of the largest Unix derivatives

The history of Unix began in 1965 when a few essays about a new operating system called “ Multics ” were published as part of the “Fall Joint Computer Conference” . Multics failed, but many of the concepts were later adopted for the Unix operating system , which was then further developed in countless derivatives .

The 1970s

In 1965, at the Fall Joint Computer Conference, several essays were published about a new operating system called Multics . Multics was a consortium made up of MIT , General Electric , Bell Labs and Honeywell . Was also involved IBM , in their language PL / I should be developed Multics.

But the Multics project failed (although it was developed further into the 1980s); the expectations of Multics were sometimes simply exaggerated, the hardware of this time could not handle a system of this size at a reasonable speed. As a result, Bell Labs withdrew from the project in 1969. Dennis Ritchie , one of the later creators of Unix and a participant in the Multics project, said in an article published in 1979:

"... The problem was the increasing obviousness of the failure of Multics to deliver promptly any sort of usable system ..."

"... The problem was the increasing obviousness that Multics would fail to deliver some form of usable system in the foreseeable future ..."

Ken Thompson (left) and Dennis Ritchie (right)

Ritchie's team, including Ken Thompson , Douglas McIlroy and Joseph Ossanna , did not give up. Above all, it was important for them to have a multi-user system that not only allowed them to program together, but also to create a real community on the basis of this system. For this purpose, the system had to support some technical specialties that were by no means taken for granted at the time, for example that several users could work on files at the same time without getting in each other's way.

While the team tried in vain to convince Bell Labs to buy a suitable machine, the technical preparatory work began at the same time: the concepts for a file system were developed on notes and blackboards , which would later become one of the core pieces of Unix. Thompson also developed some prototypes of a filesystem and primitive kernel that would run on a GE-645 . He had to stop the project for the time being after it became clear that the GE-645 would be removed from the laboratories for the foreseeable future.

He finally found a largely unused PDP-7 to which he wanted to port a game called Space Travel , which he had previously developed for Multics and GECOS . The company turned out to be more complicated than initially thought, as there was no separate development system for the PDP-7 and so all development had to take place under GECOS, which then produced the PDP-7 code.

In order to remedy this situation, Thompson began, with the help of Ritchie, the implementation of the previously designed file system, including a primitive process management system and then a series of smaller programs that were supposed to make the system usable: the editor ed , smaller file management programs , and a simple command line interpreter, later the Unix shell sh until the system was finally equipped enough to develop directly on the PDP-7 without the GECOS detour . Ken Thompson himself described the development of the operating system as a side effect of the development of the hierarchical file system , which besides the shell at Multics had fascinated him most. But in order to be able to test this file system, he had to write programs that were supposed to map the most realistic possible scenario for its use. This test program was supposed to be interactive, like an operating system, and Thompson began developing one for this purpose.

The PDP-7 was already a discontinued model at that time and did not even belong to the team, but it ultimately turned out to be an advantage, as the operating system was designed to use less resources. The financing of their own computer, which the team wanted for operating system development, was not in the interests of Bell Labs, which, after the failure of Multics, distanced itself from developing another operating system. Instead, however, the company needed a modern word processing system for its own patent office - since Thompson had already written a simple editor for Unix on the PDP-7 , the team now applied for their own computer for further development. The desired word processing system was mentioned in the application, but also the operating system required for it, which was also to be developed. Bell Labs accepted the offer and initiated the procurement of a PDP-11 . Unix was quickly ported to it and was used successfully at Bell Labs from 1971. The entire system was due to the circumstances, surprisingly small compared to today's operating systems: It consisted of 16  kB memory for the system, 8 kB for user programs and a 512-kB - hard drive . Files could have a maximum size of 64 kB. As Ken Thompson later admitted, there was also the advantage here of having developed on a weaker, but autonomous system ( PDP-7 and PDP-11 ) than on the originally desired stronger PDP-10 , which was set up as a mainframe .

In 1970 Peter Neumann coined the project name "Unics": UNiplexed Information and Computing Service . Brian W. Kernighan made a slightly mocking allusion to multics, since Unics only supported up to two users: " Emasculated Multics is Unics " ( German  Entmanntes Multics is Unics ). The name was later to be shortened to Unix. The spelling U NIX was created in 1974, according to Ritchie out of sheer enthusiasm for small caps : “ … we had a new typesetter and troff had just been invented and we were intoxicated by being able to produce small caps.

The effort in the patent office gave the group enough credibility that Unix became an interesting project for Bell Labs and to justify the purchase of a PDP-11 /74 , and the AT&T Unix Systems Group (later Unix Systems Group ) became an official project of the Bell Labs founded.

The port to C

At the same time, a development began that was decisive for the later success of Unix: the development of the C programming  language .

Ritchie and Thompson developed an interpreted programming language for the PDP-7 called  B in 1971 , which was based on BCPL . Ritchie added data types to the language on the PDP-11 , first named it “NB” (for New B ), and began developing a compiler for it.

In 1972, the re-creation of Unix began in this language, which has since been given the name C, in order to make it easier to port Unix to new computers in the future. The porting was completed in 1973 and was named "U NIX V4", ie version 4 of UNIX.


At the same time, at the suggestion of Douglas McIlroy , Unix was expanded to include the concept of pipes . Pipes connect small programs and allow the result of a program to be processed further in another program within the framework of a single command line statement, and later represented one of the important core elements of Unix, as they made the concept of small, specialized tools possible that do exactly one task.

Unix leaves Bell Labs

Unix was now running on more than 25 computers within Bell Labs, and it was first known outside Bell Labs through a lecture at an ACM symposium in 1973. The lecture was then published in a revised version in 1974 as The UNIX Time-Sharing System in Communications of the ACM . Interest in Unix also increased tremendously outside of Bell Labs.

The 1956 Consent Decree forbade AT&T , the parent company of Bell Labs, from entering new markets such as the computer market. For this reason, Unix (in the 1975 current version 6) was made available to various universities for only the price of the data carrier , together with the complete source code . In 1976, Australian professor John Lions wrote extensive comments on the source code of UNIX V6, which became famous as the Lions Book .

Unix was particularly popular at the University of California at Berkeley , where a number of improvements to Unix were immediately developed. When Ken Thompson took up a visiting professorship in the newly founded computer science department of Berkeley in 1976, the university finally became one of the most important centers of Unix development. The university later made important contributions to Unix, such as support for TCP / IP , which was later incorporated into the official UNIX version of AT&T.

From 1977 the university published its own Unix distribution under the direction of Bill Joy : the Berkeley Software Distribution (BSD).

In 1978 more than 600 computers were running UNIX operating systems.

In 1979 AT&T published the last version of UNIX with free source code, UNIX V7. This version represents a turning point in the history of UNIX as AT&T first attempted to commercialize it on a larger scale.

Microsoft acquired a Unix license in 1979 and began working on ports under the name Xenix to, among others, Intel 8086 , Motorola 68000 and Zilog Z8000 processors. On the basis of Xenix, Siemens created the first German Unix version for Intel 80186 CPUs under the name Sinix in 1984 .

In the late 1980s Microsoft co-developed OS / 2 with IBM (later split into IBM's OS / 2 and Microsoft's Windows NT ). Since OS / 2 (or Windows NT) and Xenix would have offered two server operating systems that would have rivaled each other, Microsoft decided in 1987 to transfer the rights to Xenix to Santa Cruz Operations (SCO for short, later called Tarantella ) for sale, which had been a Xenix licensee since 1983.

The 1980s

In 1980 the first port of UNIX V7 to a 32-bit machine, the VAX , with UNIX 32V and 3BSD appeared.

During the 1980s, UNIX V8, V9 and V10 were still being developed, but only presented at a few universities, although written descriptions of the work on these versions were also created. The considerations during this work ultimately also led to the development of the Plan 9 operating system .

The 1980s are marked by the beginning of the great " Unix Wars " and the commercialization of Unix. AT&T officially entered the computer market and began marketing a system based on UNIX V7, called System V , in 1983, while the University of Berkeley was simultaneously releasing 4.2BSD, which brought innovations such as TCP / IP and reliable signals with it. In the meantime, DARPA also showed interest in Unix and from then on supported the developments in Berkeley financially.

To prevent further splitting, the POSIX standardization project was launched to define a uniform interface for Unix. In 1988 POSIX.1 was finally published, a standard that was later adopted as the IEEE standard with the number 1003.1.

With NeXTStep , another commercial Unix variant was in development from 1988, which used the DARPA-financed Mach kernel in addition to BSD .

A number of (sometimes changing) alliances began to form, favoring different Unix versions:

  • Open Software Foundation : The OSF was founded in 1985, partly because of the opinion of those involved that the POSIX standard would favor AT&T too much, partly also because of fears that AT&T and Sun Microsystems , who have been cooperating since 1987, will divide the market among themselves could. Founding members of the OSF were among others DEC , Siemens , HP and IBM . The consortium had set itself the goal ofpublishinga common Unix under the name OSF / 1 .
  • Unix International : UI was founded as a direct reaction to the OSF by supporters of the AT&T line, such as Olivetti , Unisys and AT&T and Sun Microsystems .
  • X / Open : Originally founded in 1983 under the name Bison by a number of European companies such as Bull, Siemens, Olivetti, in order to be able to better represent common European interests against US companies, the consortium was later renamed with the addition of US companies .

While OSF / 1 was not finalized until the 1990s, AT&T and UI released further improvements to UNIX System V. The differences between the OSF line and UI were further publicly emphasized to the market, but the System V code bridged with it every further version the differences: with " UNIX System V Release 4" (SVR4) in 1989 practically all important innovations from BSD and Xenix were taken over into System V.

In 1987, the American computer scientist Professor Andrew S. Tanenbaum, who teaches in Amsterdam, developed a Unix-like operating system called Minix . Minix was intended to be an educational tool to demonstrate the basics of an operating system to his students, as the increasingly restrictive AT&T licenses for Unix hampered his work. Minix itself never gained much popularity, but it did inspire Linus Torvalds to work on the Linux kernel .

System V Release 4 from 1989 was later followed by Release 4.2 and Release 5.

In 1983 Richard Stallman , angry about the proprietaryization of Unix, began working on his own Unix operating system called GNU, and with the publication of the GNU Manifesto in 1985, he started a growing movement for free software .

The 1990s

In 1990 "4.3BSD Reno" was released and Sun Microsystems released Solaris 1.0. NeXTStep from NeXT Computer was continuously developed, but was initially available exclusively on NeXT hardware.

On October 5th, 1991 Linus Torvalds presented his Linux kernel with the version number 0.02. 4.3BSD NET / 2 was released. A group of BSD developers left Berkeley University and founded Berkeley Software Design ( BSDI ).

In 1992 Billy Jolitz released 386BSD , a port from 4.3BSD NET / 2 to the Intel i386 processor.

In 1993 the version "4.4BSD" appeared and BSDI published with BSD / OS a commercial version and further development of 386BSD - also based on 386BSD the development of NetBSD and FreeBSD began . Sun Microsystems published the first version of Solaris for the Intel i386 processor in 1993 and NeXTSTEP was ported to other architectures and sold separately from NeXT hardware from 1993 onwards.

On June 16, 1993, Novell bought Unix System Laboratories (USL) from AT&T .

X / Open acquired the exclusive right to use the UNIX trademark from Novell in 1993. The goal was a manufacturer-independent standard, the Single UNIX Specification (SUS). Compliance with them is a prerequisite for licensing the designation “UNIX”, which thus acquired a new meaning.

In 1994, after copyright disputes between Novell, USL and BSDI, all source code based on USL developments was removed from 4.4BSD, and 4.4BSD-Lite was released. The free BSDs also brought out new versions based on them.

There were also so-called desktop wars . See: Common Open Software Environment (COSE): HP, IBM, SCO, Sunsoft, Univel / Novell, USL → CDE , Motif , Wabi, Looking Glass.

In 1995 Novell sold its UNIX business to SCO as part of a long-term cooperation, but apart from the trademarks, retained intellectual property, including copyrights and patents.

In 1996 OSF and X / Open merged to form the Open Group . The Open Group became the owner of the UNIX trademark . In the same year, a spin-off from NetBSD appeared with OpenBSD and OPENSTEP was published, a now open platform that was jointly developed from NeXTSTEP by NeXT and Sun Microsystems.

In 1997 the Open Group published the "Single UNIX Spec / 2" standard, which now supports real-time , threads and 64-bit processors . Also in 1997 Apple Computer took over NeXT and developed OPENSTEP, with parts of Mac OS , to Mac OS X until 2001 , which however runs exclusively on Apple hardware.

In 1999, the sources of the basic operating system of Mac OS X, which like NeXTSTEP / OPENSTEP and Rhapsody are largely based on BSD-Unix, were published for the first time under the name " Darwin " under a free license.

From 2000

In May 2001, Caldera , a provider of a little-known Linux distribution, bought the Unix division of Santa Cruz Operations (SCO). SCO then renamed itself Tarantella . In August 2002, Caldera changed its name to The SCO Group , as the brand name SCO was better known than its own.

2002: ISO / IEC 9945: 2002 (Single Unix Spec)

From 2003 to 2007 lawsuits by the SCO Group against IBM and some of its customers caused a sensation: IBM had incorporated SCO patents, core technologies and copyrighted components of Unix into the Linux kernel. The allegations proved to be untenable in court, however, in particular SCO had not been the owner of the copyright of the Unix code since 1995, as claimed.

More on this here: SCO against Linux

See also


Web links

Individual evidence

  1. ^ List of articles
  2. Dennis M. Ritchie: The Evolution of the Unix Time-sharing System *
  3. Peter Seibel: Coders at Work . 1st edition. Apress, 2009, ISBN 978-1-4302-1948-4 , pp. 464 (English): Seibel: And you just wanted to play around with writing a file system? At that point you weren't planning to write an OS? Thompson: No, it was just a file system.
  4. Peter Seibel: Coders at Work . 1st edition. Apress, 2009, ISBN 978-1-4302-1948-4 , pp. 462 (English): Seibel: Do you think you benefited to being constrained by the less powerful machine? Thompson: There was certainly a benefit that it was small and efficient. … Computing going autonomous instead of centralized was, I think the really serendipitous part of it. And that rode in on the PDP-11.
  5. ^ Peter H. Salus: A Quarter Century of UNIX . Addison-Wesley, 1994, ISBN 0-201-54777-5 , p. 9.
  6. Ken Thompson: Users' Reference to B. (No longer available online.) Archived from the original on July 6, 2006 ; accessed on October 12, 2018 .
  7. GNU Manifesto
  8. ^ Report on the judgment at Groklaw
  9. ^ Report in the Heise Newsticker
  10. Application SCO against IBM (English)