Internet History: Enhancing Interactivity
Other articles in the series:
- Relay history
- The history of electronic computers
- Transistor history
- Internet history
In the early 1960s, interactive computers, starting from the delicate sprouts cultivated in the Lincoln and MIT laboratories, gradually began to spread everywhere, and in two different senses. Firstly, the computers themselves pulled out their antennae, reaching neighboring buildings, campuses and cities, which allowed users to interact with them from a distance, with several users simultaneously. These new time-sharing systems have flourished, eventually turning into platforms for the first virtual, online communities. Secondly, seeds of interactivity spread across all states, and rooted in California. And for this first seedling, one man was responsible, a psychologist named Joseph Karl Robnett Liklider .
Joseph "Apple Seed" *
* An allusion to the American folk character nicknamed Johnny Appleseed , or "Johnny the apple kernel", famous for the active seedlings of apple trees in the Midwestern United States (apple seed - apple kernel) / approx. perev.
Joseph Karl Robnett Liklider - for friends "Lik" - specialized in psychoacoustics , an area that connects imaginary states of consciousness, measured by instruments psychology and sound physics. We briefly mentioned him earlier - he was a consultant at the FCC hearing about the Hush-a-Phone in the 1950s. He honed his skills at the Harvard Psychoacoustic Laboratory during the war, developing technologies that improved the audibility of radio broadcasts in noisy bombers.
Joseph Karl Robnett Liklider, aka Lick
Like many American scientists of his generation, he discovered ways to combine his interests with military needs after the war, but not because he was very interested in weapons or state defense. There were only two major civilian sources of funding for research - these were private institutions founded by industrial giants at the turn of the century: the Rockefeller Foundation and the Carnegie Institute. National health institutes had only a few million dollars, and the National Science Foundation was founded only in 1950, and had the same modest budget. In the 1950s, in order to receive funding for interesting scientific and technical projects, it was best to count on the Ministry of Defense.
Therefore, in the 1950s, Lick entered the MIT Acoustic Laboratory, run by physicists Leo Beranek and Richard Bolt, and received almost all of the funding from the U.S. Navy. After that, his experience of connecting human senses with electronic equipment made him the first candidate for a new MIT air defense project. Participating in the Charles Project development team involved with the implementation of the Wally committee’s air defense report, Lick insisted on including the human factor in the research project, and as a result he was appointed one of the radar display design directors at Lincoln’s laboratory.
There, at some point in the mid-1950s, he crossed paths with Wes Clark and TX-2, and immediately became infected with computer interactivity. He was fascinated by the idea of complete control over a powerful machine that could instantly solve any task assigned to it. He began to develop the idea of creating a “symbiosis of man and machine,” a partnership between man and computer, capable of enhancing a person’s intellectual power in the same way that industrial machines enhance his physical abilities (it is worth noting that Lick considered this stage to be intermediate, and that later computers will learn to think independently). He noticed that 85% of his working time
... was mainly devoted to clerical or mechanical actions: searching, computing, drawing, transforming, determining the logical or dynamic consequences of a set of assumptions or hypotheses, preparing for a decision. Moreover, my choices of what is worth and what is not worth a try were shamefully determined by the arguments of clerical opportunity, not intellectual ability. Operations that consume most of the time, ostensibly devoted to technical thought, machines could perform better than people.
The general concept did not go far from the “ Memexa ” described by Vanivar Bush , an intelligent amplifier, the circuit of which he outlined in 1945 in the book “How Can We Think,” although instead of a mixture of electromechanical and electronic components, like Bush, we came to purely electronic digital computers. Such a computer would use its incredible speed to help in the clerical work associated with any scientific or technical project. People would be able to get rid of this monotonous work and spend all their attention on hypothesis formation, model building and computer goals. Such a partnership would provide incredible benefits for both researchers and national defense, and would help American scientists to overtake the Soviet ones.
Memex by Vanivar Bush, an early concept of an automatic information extraction system that complements intelligence
Shortly after this landmark meeting, Lick brought his interest in interactive computers with him to a new job at a consulting firm run by his old colleagues, Bolt and Beranek. They worked for years as consultants in parallel with academic work in the field of physics; for example, studied the acoustics of a cinema in Hoboken (New Jersey). The acoustic analysis of the new UN building in New York provided them with a large flow of orders, so they decided to leave MIT and consult all the time. They were soon joined by a third partner, architect Robert Newman, and they called themselves “Bolt, Beranek and Newman” (BBN). By 1957, they had grown to a mid-sized firm with several dozen employees, and Beranek decided that they risk saturating the market for acoustic research.
And he, of course, tracked down Liklider's old colleague, and hired him on generous terms as the new vice president of psychoacoustics. However, Beranek did not take into account Lik's wild enthusiasm for interactive computing. Instead of an expert in psychoacoustics, he received not just a computer expert, but a computer preacher who was eager to open his eyes to others. Over the course of a year, he convinced Beranek to shell out tens of thousands of dollars to buy a computer, a small, weak device, the LGP-30, made by Librascope, a defense contractor. Having no experience in engineering, he brought in another SAGE veteran, Edward Fredkin, to help set up this machine. Despite the fact that the computer mainly distracted Lik from his main work while he was trying to learn programming, after a year and a half he persuaded his partners to spend more money ($ 150,000, which, by today's money, is approximately $ 1.25 million) for the purchase of a more powerful computer: the latest PDP-1 manufactured by DEC. Lick convinced BBN that digital computers were the future, and that somehow someday their investment in experience in this area would pay off.
Shortly thereafter, Lick almost accidentally found himself in a position ideally suited to spread the culture of interactivity throughout the country, becoming the head of the new government computing department.
During the Cold War, each action had its own opposition. As the first Soviet atomic bomb led to the creation of SAGE, the first artificial Earth satellite launched by the USSR in October 1957 gave rise to a flurry of reactions in the American government. The situation was aggravated by the fact that although the USSR was four years behind the United States on the issue of a nuclear bomb explosion, it made a leap forward in missile business, ahead of the Americans in an orbit race (it turned out to be about four months).
One of the answers to the appearance of Sputnik-1 in 1958 was the creation of the Department of Advanced Research Projects of the US Department of Defense (ARPA). Unlike the modest sums allocated for the needs of civil science, ARPA received a budget of $ 520 million, three times the funding of the National Science Foundation, which itself was tripled in response to the appearance of Sputnik-1.
Despite the fact that the Office could work on a wide range of any advanced projects that the Minister of Defense considered expedient, it was originally intended to focus all attention on rocket science and space - that was Sputnik-1's decisive answer. ARPA reported directly to the Secretary of Defense, and therefore was able to rise above counterproductive and weakening the entire industry rivalry by developing a single reasonable plan for the development of the American space program. However, in fact, all his projects in this area were soon taken away by rivals: the Air Force did not intend to give control of military rocket science, and the national law on aeronautics and space, signed in July 1958, created a new civil agency that took over all issues related to space, not touching weapons. Nonetheless, after the creation of ARPA, it found reasons to survive, as it received major research projects in the areas of ballistic missile defense and nuclear test detection. However, it also became a working platform for small projects that various military agencies wanted to study. So instead of a dog, control became the tail.
The last selected project was the Orion project , a spacecraft with a nuclear-pulse engine (“explode”). ARPA stopped funding it in 1959 because it could not have been presented as anything more than a purely civilian project that fell under NASA's authority. In turn, NASA did not want to smear its purest reputation by messing with nuclear weapons. The Air Force reluctantly threw some bucks so that the project continued to develop, but he eventually died after the 1963 agreement, which prohibited the testing of nuclear weapons in the atmosphere or space. And although this idea was technically very interesting, it is hard to imagine that any government would give the green light to launch a rocket filled with thousands of nuclear bombs.
ARPA’s first intrusion into computers came about simply out of control. In 1961, the Air Force had two inactive assets on hand, which needed to be loaded with something. As they approached the deployment of the first SAGE detection centers, the Air Force hired RAND Corporation from Santa Monica, California, to train personnel and equip twenty of the small computerized air defense centers with control programs. For this work, RAND spawned a whole new entity, System Development Corporation (SDC). SDC's experience in software turned out to be valuable to the Air Force, but the SAGE project was completed and there was nothing to do. The second dormant asset was the extremely expensive extra computer AN / FSQ-32, which was requisitioned from IBM for the SAGE project, but was later recognized as unnecessary. The Ministry of Defense solved both problems,
Soon, ARPA decided to regulate this research program as part of a new information processing research unit. Around the same time, management received a new task - to create a program in the field of behavioral science. It is not clear for what reasons, but the management decided to hire Liklider as the director of both programs. Perhaps it was the idea of Gene Fubini, director of research at the Ministry of Defense, who knew Lick from working on SAGE.
Like Beranek in his day, Jack Ruin, who was then in charge of ARPA, had no idea what he was facing when he invited Lika to an interview. He believed that he was gaining a behavioral expert with some knowledge of computer science. Instead, he came across the full power of the ideas of symbiosis between a person and a computer. Lick argued that a computerized control center would need interactive computers, and therefore a breakthrough at the cutting edge of interactive computers should be the main engine of the ARPA research program. And for Lik, this meant sharing time.
Time-sharing systems appeared on the same basic principle as the Wes Clark TX series: computers should be user-friendly. But, unlike Clark, proponents of time sharing believed that one person could not effectively use an entire computer. The researcher can sit for several minutes, studying the output of the program, before making a small change in it and restarting it. And in this interval, the computer will have nothing to do, its greatest power will be idle, and it will be expensive. Even the intervals between keystrokes of hundreds of milliseconds seemed like huge abysses of wasted computer time, for which thousands of calculations could be performed.
All this computing power may not be idle if it can be shared between many users. Dividing the computer’s attention so that it serves each user in turn, the computer developer could kill two birds with one stone - to provide the illusion of an interactive computer that is completely under the user's control, without wasting most of the computing capacity of expensive equipment.
This concept was laid back in SAGE, which could serve dozens of different operators at the same time, and each of them tracked its sector of airspace. Meeting with Clark, Lick immediately saw the potential of combining user separation in SAGE with the interactive freedom of TX-0 and TX-2 to create a new, powerful mixture, which formed the basis of his propaganda of the symbiosis of man and computer that he presented to the Ministry of Defense in 1957 " A truly wise system, or Forward, to the hybrid thinking systems of the machine / person ”[sage eng. - sage / approx. transl.]. In this work, he described a computer system for scientists, very similar to SAGE in structure, with input through a light gun,
However, Lick himself did not have the engineering skills to develop or create such a system. He learned the basics of programming from BBN, but that was the limit of his possibilities. The first person to put into practice the theory of time sharing was John McCarthy, a mathematician at MIT. McCarthy needed constant access to a computer to create tools and models for manipulating mathematical logic - in his opinion, the first steps to artificial intelligence. In 1959, he made a prototype consisting of an interactive module screwed to an IBM 704 university computer with batch processing. Ironically, the first “time-sharing device” had only one interactive console - the Flexowriter teletype.
But by the early 1960s, the MIT Engineering Department had come up with the need for active investment in interactive computing. Every student and teacher who was interested in programming sat down on computers. Batch processing of data very effectively used computer time, but it lost a lot of time for researchers - the average processing time of the task on the 704th was more than a day.
To study long-term plans to meet the growing demand for computing resources, a university committee was organized at MIT, in which time sharing advocates prevailed. Clark argued that the transition to interactivity does not mean time sharing. He said that from a practical point of view, time sharing means abandoning interactive video displays and real-time interaction - and these were critical aspects of the project on which he worked at the MIT Biophysics Laboratory. But on a more fundamental level, it seems that Clark had a deep philosophical rejection of the idea of sharing his workplace. Until 1990, he refused to connect his computer to the Internet, claiming that networks were a “mistake” and they “didn't work”.
He and his students formed a “subculture”, a tiny process within the already eccentric academic culture of interactive computers. However, their arguments in favor of small workstations that do not need to be shared with anyone did not convince their colleagues. Given the cost of even the smallest standalone computer at the time, this approach from the point of view of other engineers seemed economically unreasonable. In addition, the majority at that time believed that computers - the intelligent power plants of the advancing information era - would benefit from economies of scale, just like power plants received it. In the spring of 1961, the committee’s final report authorized the creation of large time-sharing systems as part of the development of MIT.
By that time, Fernando Corbato, for his colleagues, Corby, was already working on increasing the scale of the McCarthy experiment. He was a physicist by training, and found out about computers while working at Whirlwind in 1951, while still a graduate student at MIT (the only surviving participant in this story - he was 92 in January 2019). After defending his doctorate, he became an administrator at the newly formed MIT computing center, built on the basis of IBM 704. Corbato with a team (initially it was Marge Mervyn and Bob Daily, the two best programmers of the center) called their time sharing system CTSS (Compatible Time-Sharing System, " compatible time-sharing system ”) - because it could work simultaneously with the normal process of the 704th, automatically picking up computer cycles for users as needed.
By the end of 1961, CTSS could support four terminals. By 1963, MIT had deployed two CTSS instances on IBM 7094 transistor machines worth 3.5 million each, about 10 times the previous 704's in memory capacity and processor power. The control software sorted active users in a circle, serving each of them for a split second, and then moving on to the next. Users could save the program and data for later use in their own password protected area of the disk space.
Corbato in his signature bow tie in a computer room with IBM 7094
Corby explains the time-sharing workflow, including a two-level queue, in a 1963 TV show
Each computer could serve approximately 20 terminals. This was enough not only to support a couple of small terminal rooms, but also to spread access to a computer throughout Cambridge. Corby and other critical engineers had their own terminals in the office, and from some point on, MIT began providing home terminals with technical staff so that they could work with the system after hours, without having to go to work. All early terminals consisted of a reworked typewriter capable of reading data and outputting them over a telephone line, and perforated continuous feed paper. Modems connected the terminals by telephone to a private switch in the MIT, through which they could communicate with the CTSS computer. The computer thus lengthened its senses through the telephone and signals, from digital to analog and vice versa. This was the first step in integrating computers with a telecommunications network. Integration was facilitated by the controversial AT&T status in terms of regulatory rules. The backbone of the network was still regulated, and the company was required to provide leased lines at fixed rates, but several FCC decisions eroded the company's control over the periphery, and it could hardly object to connecting various devices to its lines. Therefore, MIT did not require terminal permissions. but several FCC decisions eroded the company's control over the periphery, and that little could object to the connection of various devices to its lines. Therefore, MIT did not require terminal permissions. but several FCC decisions eroded the company's control over the periphery, and that little could object to the connection of various devices to its lines. Therefore, MIT did not require terminal permissions.
Typical mid-1960s computer terminal: IBM 2741.
The final goal of Liklider, McCarthy and Corbato was to increase the availability of computing power for individual researchers. They chose the means and time sharing for economic reasons: no one could imagine such that for each researcher at MIT they would buy their own computer. However, such a choice led to the appearance of unintended side effects that could not have been recognized in Clark's paradigm of "one person, one computer." A common file system and cross-references to user accounts allowed them to share with each other, work together and complement each other's work. In 1965, Noel Morris and Tom van Vleck speeded up collaboration and communication by creating the MAIL program, which allowed users to exchange messages. When a user sent a message, the program assigned it to a special file-mailbox in the file area of the recipient. If this file was not empty, the LOGIN program displayed the message “YOU HAVE AN MAIL”. The contents of the machine turned into an expression of the actions of the user community, and this social aspect of time sharing at MIT began to be valued as highly as the original idea of interactive use of the computer.
Lick, having accepted the ARPA proposal and resigned from BBN to head the new ARPA division, the Information Processing Techniques Office (IPTO) in 1962, quickly took up what he promised: concentrating the company's research efforts in computing to spread and improve hardware and software for time sharing. He abandoned the usual practice of processing research proposals, which were supposed to come to his desk, and he went to the "fields" himself, urging engineers to create research proposals that he would like to approve.
His first step was to reconfigure the existing SDC Command Center research project in Santa Monica. A team came from Lika’s office in SDC to reduce efforts to direct these studies and concentrate them on turning the extra SAGE computer into a time-sharing system. Lick believed that first you need to lay the foundation in the form of human-machine interaction with time sharing, and only command centers will appear later. The fact that such a prioritization coincided with his philosophical interests was only a fluke. Jules Schwartz, a veteran of the SAGE project, was developing a new time-sharing system. Like her contemporary CTSS, she has become a virtual meeting place,
DIAL 9 THIS IS JOHN JONES, I NEED 20K IN ORDER TO LOAD MY PROG
FROM 9 WE CAN GET YOU ON IN 5 MINUTES.
FROM 9 GO AHEAD AND LOAD
DIAL 9 THIS IS JON JONES I NEED 20K TO START PROGS
FROM 9 WE CAN GIVE YOU THEM AFTER 5 MINUTES
FROM 9 FORWARD START
Then, to provide funding for future projects for the development of time sharing at MIT, Liklider found Robert Fano to manage his flagship project: Project MAC, which lived until the 1970s (there were many MAC decodings - “mathematics and calculations”, “computer with multiple access” , “Machine Knowledge” (Mathematics And Computation, Multiple-Access Computer, Machine-Aided Cognition)). Although the developers hoped that the new system would be able to support at least 200 simultaneous users, they did not take into account the ever-growing complexity of user software, which easily absorbed all the improvements in speed and efficiency of iron. After launching at MIT in 1969, the system could support about 60 users using its two central processors, which was approximately equal to the number of users per processor in CTSS. However, the total number of users was much higher than the maximum possible load - in June 1970, 408 users were already registered.
The project’s system software called Multics boasted some serious improvements, some of which are still considered advanced in today's operating systems: a hierarchical file system of the tree structure with folders that can contain other folders; separation of command execution from the user and from the system at the iron level; dynamic linking of programs with loading of program modules during execution as necessary; the ability to add or remove CPUs, memory banks or disks without turning off the system. Ken Thompson and Dennis Ritchie, programmers at the Multics project, later they created the Unix OS (the name of which refers to the predecessor) to transfer some of these concepts to simpler and smaller computer systems [The name “UNIX” (originally “Unics”) was formed from “Multics”. The letter U in the name UNIX meant “Uniplexed” (“monosyllabic”) as opposed to the word “Multiplexed” (integrated), which underlie the name of the Multics system, in order to emphasize the attempt of the creators of UNIX to move away from the complexities of the Multics system to develop simpler and more efficient approach.].
Lick threw the last seed at Berkeley, at the University of California. The Genie12 project, launched in 1963, spawned the Berkeley Timesharing System, a smaller, commercially oriented copy of Project MAC. Although nominally managed by several university professors, the student Mel Peirtle actually led the work, while other students helped him - in particular, Chuck Tucker, Peter Deutsch and Butler Lampson. Some of them were already infected with the interactivity virus in Cambridge before they got to Berkeley. Deutsch, the son of a professor of physics at MIT and a fan of building prototypes of computers, was a teenager who implemented the Lisp programming language on Digital PDP-1 before he even became a student at Berkeley. Lampson programmed at PDP-1 on a Cambridge electron accelerator, as a student at Harvard.
SDS integrated Berkeley software into its new project, SDS 940. It became one of the most popular time-sharing computer systems in the late 1960s. Tymshare and Comshare, which commercialized time-sharing by selling remote computing services, bought dozens of SDS 940s. Payrtle and the team also decided to try their hand at the commercial market and founded Berkeley Computer Corporation (BCC) in 1968, but during the recession of 1969-1970, it filed for bankruptcy. Much of Pyrtle’s team ended up at Xerox’s Palo Alto Research Center (PARC), where Tucker, Deutsch, and Lampson contributed to iconic projects, including Alto's personal workstation, LANs, and a laser printer.
Mel Peartle (center) next to the Berkeley Timesharing System
Of course, not every time-sharing project from the 1960s was created thanks to Liklider. News of what was happening at MIT and Lincoln’s laboratories was disseminated through technical literature, conferences, academic acquaintances, and employees moving from one job to another. Thanks to these channels, other seeds attracted by the wind are rooted. At the University of Illinois, Don Bitzer sold his PLATO system to the Department of Defense, which was supposed to reduce the cost of technical training for military personnel. The Clifford Show created the Air Force-funded JOHNNIAC Open Shop System (JOSS), designed to increase the ability of RAND staff to quickly conduct numerical analysis. The Dartmouth time-sharing system was directly related to the events at MIT, but otherwise it was an absolutely unique project,
By the mid-1960s, time sharing had not yet completely captured the ecosystem of computers. Traditional batch processing enterprises dominated both in sales and in popularity, especially outside of university campuses. But it still found its niche.
In the summer of 1964, about two years after arriving at ARPA, Liklider again changed his job, this time moving to an IBM research center north of New York. Shocked by the loss of the Project MAC contract in favor of a rival computer manufacturer, General Electric, after many years of good relations with MIT, Lick had to share his first-hand experience with IBM in a trend that seemed to pass by the company. For Lik, a new job offered the opportunity to turn the last bastion of the usual batch processing of data into a new faith of interactivity (but it did not grow together - Lika was pushed into the background, and his wife suffered, isolated in the wilderness in Yorktown Heights. He transferred to the Cambridge office of IBM, and then in 1967 returned to MIT to head the Project MAC).
In place of the head of the IPTO, he was replaced by Ivan Sutherland, a young computer graphics expert who, in turn, was replaced by Robert Taylor in 1966. Lick's 1960 work, The Symbiosis of Man and Machine, turned Taylor into an adherent of interactive computers, and on Lick’s recommendation, he joined ARPA after working a little on a research program at NASA. His personality and experience made him look more like Lika than Sutherland. A psychologist by training, having no technical knowledge in the field of computers, he compensated for their lack of enthusiasm and confident leadership.
Once, when Taylor was in his office, an idea came to the recently appointed head of IPTO. He sat at a table with three different terminals that allowed him to connect to three ARPA-funded time sharing systems located in Cambridge, Berkeley, and Santa Monica. Moreover, they were not connected with each other - in order to transfer information from one system to another, he had to do it himself, physically, using his body and mind.
The seeds thrown by Lyclider bore fruit. He created the social community of IPTO employees, sprawling into many other computer centers, each of which formed a small community of computer experts gathered around the hearth of the computer with time sharing. Taylor thought it was time to link these centers together. Their separate social and technical structures, being connected, will be able to form a kind of superorganism, the rhizomes of which spread throughout the continent, reproducing the social advantages of dividing time on a higher level. And with this thought, the technical and political fights began, which led to the creation of ARPANET.
What else to read
- Richard J. Barber Associates, The Advanced Research Projects Agency, 1958-1974 (1975)
- Katie Hafner and Matthew Lyon, Where Wizards Stay Up Late: The Origins of the Internet (1996)
- Severo M. Ornstein, Computing in the Middle Ages: A View From the Trenches, 1955-1983 (2002)
- M. Mitchell Waldrop, The Dream Machine: JCR Licklider and the Revolution That Made Computing Personal (2001)