Maurice Wilks and Tom Kilburn

Despite the fact that the concept of a program stored in computer memory was advanced by John Eckert in January 1944, and in the same year he and John Mauchly began work on the creation of EDVAC (computer with a stored program), yet the first machines with a stored program were launched in England , at the University of Cambridge and Manchester.

Maurice Wilkes, a professor at Cambridge, who attended an EDVAC car lecture course at Moore’s school, returned to England to start developing an EDSAC computer. It was in November 1946 and finished in 1949, for two years, Maurice Wilkes, before EDVAC was launched in the United States.

And at the University of Manchester, Tom Kilburn, along with Jeffrey Tutillo, creates a computer with a stored program called the Small-Scale Experimental-Machine, or Baby for short, even earlier - in June 1948. But not only the creation of computers with a stored program, these two cores of computer science glorified their names, later they put forward a number of ideas and principles that are successfully applied in our time.

Tom Kilburn

Maurice Vincent Wilks was born on July 26, 1913 in Dudley in Staffordshire County (England). From 1931 he studied at the Cambridge College of St. John, who graduated in 1934. Then he continued his studies at Kem Bridge University until his graduation in 1937. In 1936, he was awarded the title of Doctor of Philosophy for his work on the propagation of very long radio waves in the ionosphere.

After graduating from the university, he was appointed to a junior position - a demonstrator, which corresponds to the position of assistant professor in the United States, to an open Mathematical Laboratory. His responsibilities included: overseeing the development of a new differential analyzer and consulting at the University of Manchester for further developments in this area. During the Second World War, from 1939 to 1945, he was part of the Allied army.

Upon returning to Cambridge in September 1945, he was appointed director of the Mathematical Laboratory (later Computer Laboratory), where Wilkes worked until 1980.

In May 1946, a copy of the Neumann report on the EDVAC computer was obtained and Wilkes had the opportunity overnight to read and understand a document that described the concept of recording and storing a program in a computer. He realized right away that it was real, and from that moment he never doubted that the development of computers would go that way.

After reading the report, Wilkes accepted an invitation to attend lectures on “Theories and Methods of Designing Electronic Digital Computers” held at the Moore School of the University of Pennsylvania from July 8 to August 31, 1946. Wilkes jumped at the chance to attend these lectures, although he was late for their start, but he still managed to describe ENIAC in detail and discuss the principles of creating EDVAC. Here is what Maurice Wilkes himself writes: “The most important event in my life happened in 1946 ... I managed to take a computer training course, and he made a strong impression on me.

Nothing like this has ever happened before, and only a few people knew about the achievements of Moore’s school and other founders of computer technology. The course was attended by 28 people from 20 organizations. The main teachers were John Ma-Uchli and Presper Eckert. They were on the brink of success, creating the first electronic computer ENIAC and the principles of designing EDVAC ... Subsequent events convincingly confirmed the principles that Eckert and Mauchly taught in 1946 to those of us who were lucky enough to listen to this course. ”

After Wilks returned to Cambridge, a project was created to create a computer with a stored program, which was to be carried out by the Mathematical Laboratory. The laboratory had sufficient funds to start the project, and the next three years were spent on constructing an EDSAC (ElectronicDelay Storage Automatic Calculator - an electronic electronic calculator with memory on delay lines), which was earned on May 9, 1949.

In essence, the EDSAC machine was a direct copy of the EDVAC project, but as far as programming EDSAC was concerned, Maurice Wilkes was in some way a pioneer. Charter to encode each command using binary numbers ("ones" and "zeros"), he began to search for a more convenient way to communicate with the machine. The first result of these efforts was the new codes made up of letters and short words taken from the English language. He introduced a mnemonic, where each team was represented by one capital letter: S meant “subtraction”, T - “transmit information to memory”, Z - “stop the machine”, etc.

But perhaps more valuable innovation than the mnemonics introduced on EDSAC was the subroutine library. It must be said that programmers already knew the concept of a subroutine. Grace Hopper and her colleagues used subroutines on Howard Aiken’s Harvard machine. They had notebooks with a record of the most used subroutines, so that if necessary they should not be re-compiled.

The problem was that the addresses of the location of commands and variables of the subroutine in memory changed depending on its placement in the latter. Setting the routines to a specific memory location obviously needed automation, and this was first done in the EDSAC computer. EDSAC programmers began by writing a set of unified subroutines that formed the library. After that, it was enough to enter only a short command, so that the computer independently did all the work on setting up and placing the subroutine within the main program.

Maurice Wilkes called the mnemonic scheme for EDSAC and the library of subroutines a collecting system (in English, the assembly system - hence the word “assembler”), because it collected sequences of subroutines.

Currently, programming languages ​​in which short mnemonic names directly correspond to individual machine instructions are called assembly languages. So, Maurice Wilkes is considered the creator (1949) of one of the so-called assembly languages. The EDSAC computer was successfully used in calculations for a government nuclear research project conducted in Cambridge.

In the early 1950s, a development team led by Wilkes began creating a second version of the EDSAC computer, EDSAC-II, which was commissioned in 1958.

When designing the EDSAC-II machine, the ideas of Maurice Wilks on the construction of control systems — the ideas of microprogramming — were first embodied. It is known that one of the most important issues that must be solved when designing a computer is how to control the flow of electrical signals carrying information in a computer. Currently, there are, in principle, two methods for constructing systems or control devices.

One of the methods for constructing a control system is that it is given a “rigid”, i.e., immutable, circuit of internal connections and in this form is connected to other electrical circuits of the processor. Another approach, which makes it possible to obtain a more flexible, simpler, and in some cases cheaper control system, is that the control system is replaced with a program containing detailed instructions for operating the machine in coded form. This program is placed in a separate memory block, which is introduced into the processor.

This approach to the implementation of management functions, Maurice Wilks began to develop in 1949. After two years of research, he concluded that the best approach to designing a control system is to consider it as a matrix, or a rectangular table, in which each horizontal row of cells corresponds to one clock cycle, and each vertical column is one of the control lines signals.

With this approach, the choice of a sequence of operations is simplified and boils down to the fact that binary symbols must be put in the cells of each horizontal row that would form the right combination: for each control line, which must be turned on during a given measure, you should put a unit in the appropriate cell , and in the cells of those lines that should be turned off, write zeros.

The device, equivalent to the control matrix, is a simple storage device built from repeating elements. The content of each cell in a row determines the state of the corresponding control line during one clock cycle. The set of binary digits, forming one macroinstruction, serves now simply to select a suitable series or sequence of rows in the control memory. In other words, a macroinstruction becomes an address denoting a series.

In view of this, the construction of a control system from the task of designing an electronic device turns into a software development task. Its complexity is now not in finding the right combination of circuits with hard links for generating control signals, but in correctly identifying combinations of ones and zeros to be written into the control memory.

Wilkes drew an analogy between this approach and conventional programming and borrowed terms from the programming area to describe his ideas, adding to them in each case the prefix “micro” indicating the elementary nature of management operations. So the term “microprogramming” and the whole family related to it appeared. In particular, each row of cells in the control matrix Wilkes called microinstruction, and each sequence of rows performing one macroinstruction called a microprogram.

The scientist proposed to call a memory device for storing microprograms as microprogram memory. The principle of microprogramming facilitated the understanding of control functions, and due to the fact that complex control schemes turned out to be replaced by a matrix of repetitive memory cells, the construction of equipment was simplified. More importantly, this principle allowed the machine to give additional flexibility: it became possible to change the control system without redesigning the hardware.

Maurice Wilkes presented his ideas on microprogramming in the report “The best method for designing an automatic computer” at a conference at the University of Manchester in July 1951. The method he proposed became the basis of the microprogramming technique, which was to become popular two decades later, in the early 1970s. In the same year, 1951, Wilkes published another work. Together with two colleagues, David Wheeler and Stanley Gill, he wrote the first textbook on programming.

By the 60s, after EDSAC-II, it became clear which direction the development of computers would take. As Wilkes wrote in these years, “the first computers in a certain sense were a thing for a programmer,” and rather soon the inefficiency of such use of expensive and scarce equipment became apparent. Replaced single-program mode came multiprogram mode and time-sharing mode. “It was not a consequence of the discovery of any new principle, it just became clear that the existing technical means can be used much better than before,” wrote M. Wilkes. In the United States, Britain and the USSR, work on the creation of time-sharing systems was launched, and the first CTSS was developed by F. Corbato and R. Fano in 1963 at the Massachusetts Institute of Technology.

Then, in England in the mid-60s, Cambridge scientists, in collaboration with the firm Ferranti Ltd under the leadership of Maurice Wilks, created the Titan time-sharing system.

Since 1965, Wilkes, as a professor of computer technology, together with Charles Lang participated in the creation of computer-aided design systems based on DEC PDP-7 minicomputers. This system has been improved for 15 years.

In 1974, Maurice Wilks joined the work on the creation of the corporate network of the University of Cambridge, the so-called “Cambridge ring”, using relatively inexpensive workstations. After retiring and leaving Cambridge in 1980, he became a full-time consultant to DEC, and then a member of the scientific research planning council of the Olivetti Research Board.

Upon returning to Cambridge, he became Distinguished Professor, retired, and in 1993 he was awarded the degree of Honorary Doctor of Science. Maurice Wilkes was the first president of the British Computer Society, a member of the Royal Society, a foreign member of many academies - Spain, America, Germany, etc. In 1967, he received the Turing Prize as a pioneer in such areas as computers with a stored program, subroutine libraries and microprogramming. Our further narrative is dedicated to another pioneer - Tom Kilburn, a humble professor from Manchester.

He was born in 1921 in Dewsbury, Yorkshire (England). After high school, Kilburn entered the University of Cambridge, where mathematics became his passion. Upon graduation, he received a bachelor of science degree and then a master of science.

In 1942–1946, Tom Kilburn conducted research in the field of electricity, magnetism, and electronics at City & Guilds in London, and then worked at the Research Institute for Long-Distance Communications. Here he met with the new boss, and then colleague Frederick Williams. Tom Kilburn quickly impressed everyone as a keen-minded and technically literate electronic circuit engineer and as a member of the Williams group made many improvements in radar electronic circuits.

A significant development of the Williams group was the use of a cathode ray tube (CRT) for storing data bits. The so-called Williams tube recorded dots and dashes on the phosphor coating of the screen, which represented respectively 0 and 1 binary data. Placing a tube on the screen in a metal plate, one could get a capacitive coupling, and then use an electron beam to read these points and use a dash. The use of Williams tubes allowed arbitrary selection of stored information (the first RAM) instead of sequential sampling using mercury delay lines.

In December 1946, Frederick Williams and Tom Kilburn transferred to the University of Manchester for

Bourn transferred to the Manchester University in order to improve the design of the memory tube and at the same time begin to develop what later became the first computer with a stored program.

Tom Kilburn got acquainted with computers developed in the USA in 1945, and then studied them in more detail in early 1947, when he attended Alan Turing's lectures at the National Physical Laboratory (NFL). It is known that at this time Turing and his colleagues from the NFL were developing a computer and details of its creation were presented in Turing lectures.

Having prepared a bridgehead in this way, Kilburn, along with Jeffrey Tutill, led the development of the Baby computer (Small Scale-Experimental-Machine) with a memory on the Williams tube of 1024 bits. June 21, 1948 Tom Kilburn calculated the first program on the computer Baby. The program to determine the maximum number multiplier was implemented by a computer in 52 minutes.

The news about the appearance of the Baby was picked up by journalists: “A miracle of our time ...”, “A memorizing machine, capable of solving the most complicated mathematical problems” ... - the headlines read. The inventors were inundated with letters from biophysicists, nuclear physicists, meteorologists with requests to make calculations, but the creators of the computer were interested in orders for the development of new machines.

In the future, Baby was finalized and became known as the “Manchester Mark-1 ″ (so as not to be confused with“ Mark-G Howard Aiken). In 1951, the company Ferranti Ltd put on the market one of the world's first commercial vehicles, which was called Ferranti Mark 1.

In the early 50s, Tom Kilburn worked at the University of Manchester as a lecturer in the electrical engineering department, and then began his unique development - the Atlas machine.

The Atlas machine is the first computer system in which many devices and principles that are currently recognized as standard, including a virtual (logical) address space, exceeding in terms of the actual (physical) address space, single-level memory on magnetic cores with duplicate a magnetic drum; and an architecture based on the use of an operating system with hardware to facilitate programming (for example, with extracodes).

Work on the creation of this machine began in 1956 under the direction of Tom Kilburn at the University of Manchester. Since 1958, the Atlas project has been funded by Ferranti Ltd. The prototype of the machine was tested in 1961, and the first industrial designs of the machine appeared on the market in 1963.

It is particularly necessary to emphasize that the first time Kilburn created the Atlas machine was to implement the concept of virtual memory, from which the method of dividing memory into pages arose and dynamic translation of addresses by hardware became possible.

Atlas machine had a high performance - about 900 thousand op / s, which was achieved by improving the principles of multiprogramming and the use of high-quality transistors. Some scientists consider Tom Kilburn and his fellow developers of the Atlas machine the founders of modern operating systems (OS), which have defined the functions of the OS and the importance that they will acquire in the computers of the next decades.

After completing work on Atlas in 1964, Kilburn headed the first computer science department in the British Isles, established at the University of Manchester. After receiving a doctorate in 1953 for research in the field of computer technology, Tom Kilburn worked at the University of Manchester until his retirement in 1981. For services to the country in 1973, he was awarded the rank of captain II rank.

In June 1998, the University of Manchester solemnly celebrated the fiftieth anniversary of the first computer with the stored program. In honor of this anniversary, a group of engineers decided to assemble an exact copy of Baby. For several months in a row, enthusiasts searched for old vacuum tubes and even found metal racks for Baby - they served as a hedge in some backyard. And ICL PLC invited enthusiastic programmers from around the world to take part in a competition to create applications for a recreated computer. They also remembered Tom Kilburn, who was asked to lead the jury of this contest. A modest professor from Manchester, who turned 77 years old on his jubilee, lived in a standard English house, where “the greatest wonders of civilization were a dishwasher, a TV and a car”.

To the question of the ubiquitous journalists about the day of today's computer industry and communications, Tom Kilburn said: “The money had nothing to do with our work. We were only interested in science, and Gates already lives in another world. ”

At the end of January 2001, a message appeared in many foreign periodicals: “Tom Kilburn, a pioneer of computerization and a participant in the creation of the world's first computer with a stored program, died last week at the age of 79 after a long illness.”