POSIX Impact

A friend and colleague Donn Terry passed this week. Donn was one of the early and key leaders in the POSIX (UNIX/LINUX) standards effort, and considered this to be one of his major contributions. I fully agree with Donn.

The IEEE POSIX work started in 1984 as an outgrowth of work initiated in the /usr/group organization to build a standard on the two variations of UNIX at the time - version 6 and BSD (Berkeley). Dr. Steve Sherman of U. Nevada had initiate a similar project in IEEE (via a "PAR" - Project Authorization Request), and Heinz Lycklama (for /usr/group) and myself (for IEEE) served as co-chairs of the initial working group. The first version of the IEEE Std 1003 was published in 1988 with the reference "IEEE IX" -- building on the popular use of "IX" suffixes on brands of the UNIX System. At the suggestion of Richard Stallman, this was changed to POSIX for "portable operating system interface".

Donn joined the effort early when he worked with HP and accepted the role of vice chair when Heinz went on to other interests. As the effort expanded, Donn suggested breaking the different topics into different documents and working groups. This created the basis for the PASC (Portable Applications Standards Committee) which I chaired and Donn took the lead on the 1003.1 operating system interface. We formed additional committees to address the shell environment (1003.2) led by Hal Jespersen and Don Cragun.

The interest of the US Federal Government NIST (previously National Bureau of Standards) at using this for government procurement, an effort led by Roger Martin, drew in participation of a wide range of corporate interests. Project 1003.3 was created to define "test methods" for the other projects. The focus unambiguous wording to facilitate testing resulted in significant improvements in the clarity of the documents, and ultimately a conformance test suite developed by NIST. (Roger chaired the 1003.3 effort and a set of real time interfaces was initiated as 1003.4 with Susan Corwin as chair.)

A meeting with the "BISON" (Bull, ICL, Siemens, Olivetti and Nixdorff) group at a meeting in Florence with the publication of that first version of the document resulted in early collaboration between the two, and initiation of the ISO POSIX work at the international level. The resulting ISO/IEC STD 8845 projects were developed in parallel with the IEEE work. Donn and the other PASC leaders were active participants in this effort, providing guidance and input at the IEEE technical meetings and the ISO/IEC SC22/WG15 POSIX working group. (BISON became X/Open became The Open Group.)

We held one of our international meetings in Sweden in conjunction with a UNIX conference there (1991?). A student participating in that had a discussion with us about his work, indicating he was developing his own operating system, which we suggested might follow the POSIX standard. The result is the LINUX system, which may well have expanded beyond the initial System V (after system 7) AT&T product and BSD versions in terms of installations.

At this point, 2022, a significant number of POSIX compatible or influenced systems exist in servers, cloud computing centers, high performance computers, Apple systems (initially BSD variations), and many cell phone systems. The number of devices measuring into the billions. There are many tales to be told about the twists and turns in the evolution of this set of standards, perhaps in another post.

Looking back at the role Donn played, along with a few other key leaders, it is clear to me that the overall set of projects would never have engaged industry, government, and the international acceptance without their personal commitment, collaborative capabilities and technical understanding. Donn's passing brings into my focus the level of teamwork and commitment that is involved in changing the world. And yes, we did change the world -- thanks to this team.