UNIX: A History and a Memoir

Today, I finished reading “Unix: A History and a Memoir” by legendary computer scientist Brian Kernighan. Kernighan’s accomplishments are too many to list in this space (writing the definitive book on C, co-author of the AWK programming language, to name just a couple), but some people may not know that he was also deeply intertwined with the development of Unix. In this short volume, Kernighan recounts a comprehensive history of the extremely influential and widely used operating system, told from a personal perspective. Unix, in its early days, was largely the product of Kernighan’s colleagues Ken Thompson and Dennis Ritchie at Bell Labs. But Kernighan was actively making contributions almost from the beginning. His intimate personal knowledge adds a lot of value to the book. Kernighan maintains a good balance between “official history” and his own involvement throughout. He explains concepts related to Unix and its ecosystem clearly and methodically. There’s no doubt that this book has a limited audience. In my opinion, to find the book interesting, you need to have a pre-existing interest in computer history, Unix, and programming (in that order). If you already have some familiarity with using Unix (or its derivatives) from the command-line, that will certainly help you understand the significance of many of the items that Kernighan discusses. If you have no prior experience with Unix, then I don’t know why you picked up this book or read this review! At just 180 pages, with plenty of illustrations, “Unix: A History and a Memoir” is an easy read. Yet, Kernighan still manages to pack plenty of detail. He concentrates the most on interesting user-facing innovations within Unix, and innovative programs that became standard pieces of its ecosystem. Kernighan explains clearly how all of the pieces fit together and evolved from one another. This provides interesting insights for software developers and system designers. Kernighan also spends plenty of pages on the human-side of Unix, including short vignettes about his colleagues and what the work environment was like at Bell Labs. I appreciated these touches and they really helped paint a complete picture of the operating system’s development in my mind. Kernighan is a good story teller. Kernighan has written many widely read technical books published by highly regarded outlets. I have previously read his books “The C Programming Language” and the “The Go Programming Language.” Like those books, the writing and editing in this self-published memoir is of the highest quality. Yet, a minor point is that the cover design is not. It’s pixelated and looks like something straight out of the ’80s (maybe he was going for that aesthetic). This is ironic given the book’s significant content on type-setting software. I almost wonder if Kernighan did this to make a point along the lines of “don’t judge a book by its cover.” More likely, he just didn’t realize it would come out that way. Even Brian Kernighan makes mistakes. “Unix: A History and a Memoir” is an excellent book that achieves the wonderful virtues of Kernighan’s other books by being succinct, comprehensive, and clear at the same time. Kernighan is a talented writer, and every word is more meaningful because he lived the subject matter inside and out. The book has a quite limited audience, but if you are in that audience, you should definitely check it out.

This book answered not only questions I had but also questions I didn’t know I had about the birth of the UNIX operating system. It opened a window into the research department of Bell Labs and revealed the supportive work environment that enabled such great strides. The UNIX operating system and its spin-off LINUX now comprise the backbone of critical computing infrastructure, the reliability of which is depended on by all of us for so much of our daily functions.

I was late to the party, joining in the mid-nineties as it were, when the state of play was much advanced so it was nice to get more familiar with the origins of my favorite toy. I was not surprised when Unix derivatives came to absolutely dominate the computing world.

"UNIX: A History and a Memoir" is a small book by Brian Kernighan... better known as the author of "the C programming language." Although it is called a memoir of Unix... it is mostly a memoir from Brian Kernighan and the wonderful environment he was part of in Bell Labs. It is that environment that led to Unix and many other contributions to our industry. The book is written from an extremely humble perspective... Brian doesn't take credit for much as he claims that it was all Dennis Ritchie and Ken Thompson. The book starts with how Brian ended up in Bell Labs, the research lab from AT&T. In the first few chapters, the book describes the culture that was created at Bell Labs. A culture of collaboration and experimentation. A bunch of geeks trying out whatever came to mind, sharing it with each other, and giving each other lots of feedback. Especially the description of the work in the Unix room, the atmosphere, the pranks, and the photos were wonderful to read. Chapters 4 and 5 become a little bit more technical where Brian explains the different parts of a Unix system and the history of that. For example, he describes the influence of Doug McIlroy on pipes. Him convincing Ken Thompson to try it out.. and Ken doing that overnight and liking the result. Or the history of grep and how its name was simply the name of the sed expression for searching and printing lines. The book is full of these stories of how and why things were created. It was wonderful to read! After chapter 5, Brian speeds up and just summarizes what happened with Unix after the creation of Unix... outside that Bell Labs environment. Although these chapters were still interesting, they weren't as nice as the Bell Labs environment and creation stories. The book ends with some speculation on how so much innovation has come out of Bell Labs at that time... and his ideas on whether and how that could be reproduced. I thoroughly enjoyed this book. I enjoy reading historical and biographical books on technology and this was a very good one. If you enjoy Unix and want to know more about its history, this is definitively one of the books to read. 4 stars!

I’ve been hoping for many years that Brian Kernighan might someday write a Unix history book. No one else on earth is as qualified to write such a book. This book touched on a lot of interesting topics, including: • The technical history of Unix (how it descended from Multics, how pipes were invented, etc.) • The origins of various tools that we take for granted today, including diff, awk, and more • Why the Unix shell language doesn’t resemble C • The origins of C and its impact • Brief biographies of several of the main players in Unix • Humorous illustrations of the laid back culture at Bell Labs • A brief history of Bell Labs and some important discoveries made there in various fields • What made Bell Labs such a ripe environment for innovation • Could something like Unix happen again? (Spoiler: His belief is that there will never be another major operating system that isn’t a descendant of Unix, but we will definitely see such innovation in other areas.) What really ties this all together is Dr. Kernighan explaining all of this from his own perspective. He explains how he got to Bell Labs, what his relationships were like with his famous colleagues, how he became interested in programming languages and writing books, and much more. As for the quality of the printing, I can say this. The cover looks like a JPG, but the text inside is as sharp as any other book. The quality of the cover doesn’t detract at all from the enjoyment I got from this book. I hope this is only the first of a series of history books from Dr. Kernighan. There’s still more to be said about the people who created Unix. I would love to see entire books on the biographies of Ken Thompson and Dennis Ritchie, and many more of the people mentioned in this book. There’s not much written about their lives before Bell Labs that I’m aware of. I would love to know what made them such innovative thinkers. Wonderful book. I’m very thankful for this.

I have trouble not giving high quality writing 5 stars. Kernighan’s writing style is a perfect fit for a memoir. I thoroughly enjoyed the anecdotes particularly the unexpected cameos and backgrounds on the origin of those commands that I type every day. If you have read Kernighan’s previous works and enjoyed them, then you will enjoy this one as well. If you haven’t read anything from him and have the slightest interest in computing history then you will likely appreciate this book as well. I did read the print version which I found the be well formatted though the binding is sub-par in my opinion. I cannot speak for the digital version. I did come close to knocking off a star for the binding quality. If 4-1/2 were an option, then this would be a 4-1/2.

It was a lot of fun reading this book about the origins of Unix and supporting programs. Brought back memories of working on an AT&t 3B2 computer. Many of the design philosophies and development processes are as relevant today as they were then . The breakup of the Bell telephone company, much like that of IBM, has caused these great research centers to end. Much of what we think of innovative software today is really no different than what it was 40 years ago. As Isaac Newton said, if we can see further today, it's because we are standing on the shoulders of these Giants. A great read well written.

Brian not only told the story to us he relived it through this book for us. I opened it up because I was a little curious and didn’t put it down until I finished it.

L'histoire d'Unix, et du context dans lequel il est apparu, racontée par l'un de ses pères. Très intéressante lecture.

Know the legends behind the legends , there are many ken & Denis behind Unix. Brian is one of that geek who enjoys writing.

Betriebssysteme werden in der Regel nicht mit Legenden oder Philosophie in Verbindung gebracht, aber Unix ist anders... – Unix hat seine Paradigmen, Helden, Mythen und treue User Groups. Brian W. Kerninghan ist Professor für Computerwissenschaft an der Universität Princeton, er ist Mitautor von Büchern wie “The C Programming Lanuage“ (mit Dennis M. Ritchie), “The UNIX Programming Environment“ (mit Rob Pike), “The AWK Programming Language“ (mit A.V. Aho und P.J. Weinberger), die einen regelrechten Kultstatus erworben haben und immer wieder neu aufgelegt werden; aus neuerer Zeit gesellt sich “The Go Programming Language“ (mit A. Donovan) hinzu. Das hier vorliegenden Büchlein ist dem Ursprung und der nunmehr fünfzigjährigen Geschichte von Unix gewidmet. Zwar war der Autor nie direkt an der Entwicklung von Unix beteiligt, doch das Computing Science Research Center von Bell Labs war der Ort des Geschehens, und sein Büro nicht weit entfernt von denen von Ken Thompson, Dennis Ritchie, Bob Morris, Joe Ossanna und German Holzmann. Da Kerninghan sich bei der Vorbereitung des Buchs nicht allein auf seine eigenen Erinnerungen verlassen wollte, führte er mit den Protagonisten – seinen ehemaligen BTL Kollegen, von den glücklicherweise noch viele am Leben sind, etliche Interviews, von den einige auch als Video verfügbar sind. Nachdem sich Bell Labs 1969 aus dem Multics Projekt zurückgezogen hatte, beschäftigt sich Ken Thompson mit Programmen zur Steuerung einer Disk an einer wenig benutzten PDP-7. Schließlich realisierte er, dass er sich nur drei Wochen von einem Time Sharing System entfernt befand, er brauchte je eine Woche für die Entwicklung eines Editor, Assemblers und eines Kernel- Overlays. Er versuchte dabei das Beste aus Multics einfließen zu lassen, diesmal aber in einem betont einfachen, aber eleganten Ansatz. Mitte 1969 war Unix als Betriebssystem soweit gereift, dass es eine kleine User Gemeinde hatte, zu der neben Ken Thompson und Dennis Ritchie, Doug McIlroy, Bob Morris, Joe Ossanna und der Autor gehörten. Der begrenzte Speicher der PDP-7, gerade 4K 18-bit Words, behindert aber bald die Entwickelung von Compilern. Ein Antrag zu Beschaffung einer PDP-11 wird zunächst abgelehnt – Bell Labs hat nach dem Multics Desaster wenig Interesse an Betriebssystemen. Joe Ossanna kommt aber auf die Idee, die Maschine für die Patentabteilung, die ein Textverarbeitungssystem benötigt, anzuschaffen. So geschah es, dass Thompsen Unix für die neue Hardware portierte. J. Saltzers runoff für CTSS wurde von D. McIlroy mit BCTL als roff für Unix umgeschrieben, und schließlich in Assembler neu implementiert. Mit roff konnte nun das Patent Office seine Patentschriften und Dokumente flexibel formatieren – nachts stand die PDP-11 den Systementwicklern für ihre Zwecke zur Verfügung. Damit wurde 1971 die 1. Edition von Unix fertig gestellt, seit dieser Zeit gehört auch das Programmer's Manual (man), in seiner noch heute verwendeten Struktur, als fester Bestandteil dazu. Unix wurde, wie andere Betriebssysteme, zu dieser Zeit üblich, in Assembler geschrieben. IBM hatte 1964 PL/I zur systemnahe Programmierung entworfen, der Compiler war aber für Unix viel zu mächtig. Thompson experimentierte mit B, das Dennis Ritchie zu C weiter entwickelte – insbesondere types kamen hinzu. Thompson unternahm 1973 drei Versuche, den Unix Kernel mit C umzuschreiben, da C wie die PDP-11 Architektur Byte orientiert war. Das war aber erst von Erfolg gekrönt, als C schließlich über strucs verfügte, diese Version floss dann in die 6. Edition von Unix ein. Doug McIlroy hatte bereits 1964 die Idee, eine Methode zu schaffen, mittels derer Programme beliebig miteinander kooperieren können. Als Thompson darüber nachdachte, fand er die ursprüngliche Graphen- artigen Beziehungsstruktur von Programmen zu komplex; als er die Sache weiterverfolgte, entwickelte er schließlich die Idee von Pipes – das Faszinierende daran war, das sich Pipes beinah trivial in den Kernel integrieren ließen. Über Nacht passten Thompson und Ritchie noch die damaligen Unix Tools an, so dass sie per default Standard Input und Output verwenden, dazu mussten lediglich zusätzliche Ausgaben auf stderr umgeleitet werden; Thompson baute den Pipe Mechanismus in die Shell ein – und fand das Ergebnis “mind-blowing“; am nächsten Morgen funktionierten all die grandiosen Einzeiler, wie 'who | grep | wc...'. Die Einführung von Pipes stellt einen gewaltigen mentalen Sprung vorwärts dar, gemeinsam mit der genial einfachen Syntax, wurde dadurch die Unix Toolbox Philosophie auf den Weg gebracht, wonach jedes Programm für nur eine Sache zuständig sein – diese aber bestmöglich erledigen – soll, den kleinsten gemeinsamen Nenner der Interoperabilität stellen dabei Textnachrichten dar. Damit waren alle wesentlichen Ingredienzien von Unix beieinander, Thompson und Ritchie erläuterten ihren Ansatz und ihre – für die damalige Zeit relative große – Implementation auf einer PDP- 11/40 in dem Paper “The UNIX Time- Sharing System“, das m Juli 1974 in den Communications of ACM erschien. Darin betonen die Verfasser, dass der Erfolg von Unix gerade auf seinem minimalistischen Design, das nur wenig vordefinierte Strukturen kennt, beruht – so kennt das das Filesystem neben regulären Files, nur directories und special device files, deren I/O aber stets einheitlich von read/write System- Calls behandelt werden, reguläre Files kennen ferner OS- seitig keinerlei zusätzlich Strukturen, diese bleiben allein der Interpretation von Anwendungs- Programmen vorbehalten. Zudem ist das API von Unix frei von komplexen Control Blöcken u.ä. Da ferner auch der Unix Kern fast vollständig in C beschreiben ist, erleichtert das die Portierung von Unix. Im Folgenden geht der Autor noch auf viele weitere Stationen der Erfolgsgeschichte von Unix ein, zunächst werden Lizenzen an Universitäten vergeben, wenig später wurde kommerziell verwendbar gemacht. Das System verbreitet sich fortan lawinenartig, einschließlich seiner verschiedenen Derivationen wie BSD, SunOS, Xenix, HP-UX, IRIX, AIX und schließlich Minix und Linux. Brian Kernighan ist mit seiner 'Unix- Geschichte' sicher ein ganz besonderes Buch gelungen, er hat sich ganz bewusst für eine nicht- technische Darstellung entschieden, wie er im Vorwort bekennt, dafür flossen aber zahlreiche persönliche Erinnerungen und Erinnerungen anderer unmittelbar Beteiligter ein, die das Buch zu einem ganz besonderen Zeitzeugnis machen. Es steht damit sicher in einer Reihe mit Peter Salus “A Quarter Century of Unix“ und komplementiert dieses in hervorragender Weise. Leider ist das Buch etwas mager ausgestattet, es fehlt ein Index, unverständlich bei einem historischen Werk, und statt einer Bibliographie gibt es nur ein Prosa-Kapitel 'Sources'.

This book feels quite short, but perhaps that's because it's a real page turner for anyone who is interested in the history of UNIX. As other reviews have mentioned, the Kindle edition is fixed format which means it doesn't autoscale very well on Fire devices etc (I found myself having to zoom in slightly on every page turn). Brian's writing is engaging, and he brings together many anecdotes, some of which UNIX enthusiasts will likely have heard elsewhere (but perhaps unconfirmed and/or uncited) into a single text that flows really well from start to finish. Despite the book being a "memoir", Kernighan also covers the modern day status quo and how we arrived here. If you can overlook the formatting/Kindle flow issues, this is highly recommended.

Let me start with a little bio of my own. I started on UNIX in 1975 on a GE mini, and over the next five decades worked on several versions of UNIX, wrote dozens of books and thousands of magazine articles on UNIX and Linux (as well as a few offshoots and competitors) and was the Technical Editor for half a dozen UNIX and Linux magazines. I still use a UNIX variant every day. One of my first books on the subject had a "history" of UNIX that was woefully short (although mostly accurate). I also love the C language, have written books and articles on it, was involved with the C User's Journal, and taught hundreds of classes in Universities and for companies. My point to this is that I came to this book with some knowledge of the background, the players involved (having interviewed many for articles decades ago), and a lot of hands-on experience. And still I learned a ton from this book. Kernighan is best known for his role in K&R C, essentially developing the language that UNIX was written in (after the first machine language version). His history of UNIX and C comes from being involved directly, and having intimate knowledge of the players and the subject. He also has written some classic books, so it's no surprise this book is well written and interesting to read. For those who are not technically literate, he explains things well; for those who are technically competent, this doesn't distract from the narrative. Technical people won't learn anything new from a technical point of view, but you will learn more about the history of Unix and the author and his colleagues. It's a first-person review of the subject, different from the other histories I've read (and written!) because this was the guy who was there. And remembers it all! The book opens with a view of Bell Labs, and the way it handled its scientists and their research. I spent many years in the Canadian equivalent, Bell Northern Research (BNR) which became Northern Telecom Labs (Nortel) as a Member of Scientific Staff, which parallels Bell Labs a great deal. From there, Kernighan looks at the development of the early time-slicing operating systems, including CTSS and Multics, and how the latter led to UNIX. This is quite possibly the most accurate and succinct history of the subject I've ever read. Along the way, he touches on many subjects, including classic computer science problems, the evolution of programming languages, the development of UNIX (which led to parallel operating systems like Xenix [Microsoft and Santa Cruz Operation] and Linux). The book is also a memoir of Kernighan's life and work, and how that life was tied to the operating system and languages he was involved in (including one of my all-time favorite scripting languages, awk). The book is an easy read, especially if you are technically versed, but even if you're not you won't be glossing over or bored. I pretty much read the book in a day, and filled some of the history I wasn't aware of, and corrected some errors in my knowledge too. As a history of both UNIX and the author, this is an excellent book, and I highly recommend it to anyone with an interest in UNIX or computer science.