"The tragedy of common sense
it that it is not
very common."
(Albert Einstein)

"Politically correct Christianity
is tolerated but despised.
Full Gospel Christianity is
respected but persecuted."

"If you marry the Zeitgeist
you will soon become widow."

"To reach the source of a river
you must swim upstreams."
(Stanislaw Jerzy Lec)

"I note that all those,
who are positive to abortion
already are born."
(Ronald Reagan)

Last modified: 2020 11 28 16:43

From the computer's infancy in Sweden — personal memories from the 1950s onwards


Let's begin with some not so accurate forecasts regarding the future development of computers:

I think there is a world market for maybe five computers.
             Thomas Watson, Chairman of the Board of IBM, 1943
There is no reason why anyone should want a computer at home.
             Ken Olsen, CEO of the mainframe company Digital Equipment, 1977
ENIAC is equipped with 18,000 electron tubes and weighs 30 tons. In the future, computers may have only 1,000 tubes and weigh only 1.5 tons.
             The magazine Popular Mechanics about the then most powerful computer and future computer development, 1949.

As the reader well knows, these bleak predictions have been exceeded by a factor of a billion or more!


(I apologize for the generally poor image quality below. But most of the images are from old brochures that were rasterized. I have tried to make them look good and reduce the raster effect but it is difficult to get them perfect. Hopefully they provide the information I want them to give. In addition, they convey a contemporary colour of how brochures at that time looked. The pictures have kindly been made available to me by Erling's daughter (and my cousin) Birgitta in London)

I dedicate this article to my uncle Erling Kaiser (1919-2000). He was my foster father from the time I was 8 until I turned 14. Erling was Danish and came to Sweden shortly after the end of World War II (WW2). There he met my aunt Vera and became smitten. Erling was thus my uncle-in-law. After having had a few different jobs, including as a mortuary caretaker (you take what you can get), he was offered a position as a clerk at the Swedish Pensions Agency (which in 1961 became part of the National Insurance Administration).

During the war, computers had begun to be used, both for calculations (e.g. to crack the German and Japanese codes) and for administration (the US war machine was a giant machine and of course required extensive and efficient administration). Since calculating pensions for millions of people requires a large calculation capacity, plus it is not particularly uplifting to do such things manually, the Swedish Pensions Agency began to think about whether this could be automated with the new machines. Since there was no one at the Swedish Pensions Agency who knew anything about computers, and Erling had proved to be both skilled, responsible and intelligent and was also young (and thus could be expected to have the ability to rather easily learn new ways of thinking), he was sent on a course. The machines purchased came from the French company Bull (Compagnie des Machines Bull), which at this time, alongside IBM, was among the first to introduce "punch card-based computers" or as they were also called "calculators". Bull is still (2020) operating internationally often under the name Groupe Bull.

It turned out that the young Dane had an aptitude for computers and programming and he, despite his young age, got involved in creating the computerised pension system in Sweden. Which he apparently did with great success. Of course, he learned a lot during this time, which came to good use when he started his own computer company a few years later.

But now we have to take a side track. Vera's (i.e. Erling's wife’s) father, Nathan Behrn (who was also my beloved grandfather), owned a large shoe store at Norra Bantorget in Stockholm (at the corner of Upplandsgatan and Barnhusgatan). Nathan had become ill with cancer and could no longer manage the business, which incurred ever-increasing debts and eventually stood on the brink of bankruptcy. Erling, who was a very good administrator and also had an entrepreneurial heart, saw his chance here and made a deal with the bank (i.e. the largest creditor). "If I make sure you get all your money back, I want to borrow a very large sum to start a private computer company" (as far as I know, Erling's company became Sweden's first, and also for a number of years Sweden's largest, privately owned computer company). The bank agreed to this and Erling was allowed to run the shoe business as he wished. Through hard work and great ability, he managed to get the shoe store on its feet within a couple of years and the bank got all its money back.

I do not remember what happened to the shoe store, maybe it was sold, maybe it was closed down. Whatever, Erling got his money and in 1954 was able to start his computer company, which was named Hålkortscentralen Siffer-Service (Hålkortscentralen means "the punch card centre" and Siffer alludes to "figure" or "digit" of "numeral"). The address was Kungsgatan 74 (in the very centre of Stockholm) and remained unchanged until Erling sold the company in connection with his retirement.

Figure 1. An advertising brochure for Siffer-Service of a later date. Probably from the mid-1960s. Today the brochure would probably have been written in English but the world was a smaller world in 1960. (Click here to see the picture in better resolution).

The word "punch card centre" refers to the fact that at this time data was stored on so-called punch cards. By punching machines holes were made at certain predetermined positions on these cards, which coded for the data that was stored. Besides punch cards there were also punch tapes, which were long, narrow paper strips with punched holes (see figure 1, the third and fourth small picture from above) — the advantage of the punch tapes was that they could store an almost unlimited amount of information while the drawback was that they easily broke). When I worked with the advanced mainframe computer at Fysikum (the physics department) at Stockholm University in the mid-1970s (a Control Data 3600 — by the standards of the time, it could be considered a supercomputer), we still used punch cards. But not to store data but to store the program code itself (which could consist of a stack of thousands of punch cards). By this time magnetic tape (and to some extent punch tape) was used for data storage. Regarding the program code, there were some advantages to having it on punch cards. If you wanted to change or correct a line of program code, all you had to do was take the current card out of the stack and insert a new card at the same place. If you wanted to add a new subroutine, you simply placed this card (or cards) last in the stack.

Figure 2. Close-up of a punch card. The card in the picture does not store data but a statement in a Fortran program (Fortran is a powerful programming language for numerical calculatione and is still used today). The program statement is also, which was standard, written in plain language in the upper, left corner — "Z(1)=Y+W(1)". In the upper right corner is stated which program project the statement is part of (PROJ039). To the left on the punch card (under the plain language text) the program statement is punched and on the right side the project information. People who worked with punch cards soon learned to read the punched information directly from the holes, but having the plain language characters printed, simplified everything.
Note that the card is chanfered (cut with an inclination) in the upper left corner (this was standard). Thanks to this you could immediately see if a card was turned the wrong way (upside down, back to front etc) in a thick stack of cards.

At Siffer-Service and similar companies there was a special department, the punching department, where there were punching operators (often ladies) who transferred various types of data (salary information, working hours, etc.) from documents to punch cards (see figure 1 the second small picture from above). This was done via card punching machines, which initially only had 12 keys (some characters were then created by pressing two keys at the same time) but eventually got complete alphanumeric keyboards. Everyone who has worked with punch cards remembers that mistakes were common in the punching process, which was a major problem, and various methods were developed to prevent and detect such errors.

Click here to watch a demonstration of a card punching machine (4:14 min).

The format of the most common punch cards was 187x83 mm and they had 80 columns (vertical rows) and were made of rigid cardboard. Each of the 80 columns could store one character, i.e. each punch card could store 80 characters (sufficient to e.g. store a person's name, address, telephone numbers and other data). The so-called EBCDIC alphabet was used for the coding, which could represent 256 different characters (click here for more details about how punch cards were coded). Each column thus contained 8 bits (because 28=2⋅2⋅2⋅2⋅2⋅2⋅2⋅2=256) and 8 bits are the same as 1 byte. In modern terminology 80 bytes could be stored on one punch card (1 byte per column and 80 columns). To store 1 Gb (1 gigabyte = one billion bytes) 12.5 million punch cards would be needed, which is a 2.2 km high stack of punch cards! For a short and easy explanation of bits and bytes click here!

Punch cards started to be used early in the 19th century to control looms. At the 1890 census in the USA data was entered on punch cards, which were then run in a primitive predecessor to the tabulator. In this machine the results were calculated. By using punch cards, the time of calculation for the census was shortened from 2 years to 3 months. At this census, they chose a format for the punch cards equal to the 1880 dollar bill (perhaps because then, after minor adjustments, you could use the mechanism in banknote counting machines to handle punch cards) and this format has since then been retained (a classic boarding card used by airlines has exacly the same format).

During the census of 1890 they were not content with just counting the US population, but they also wanted to know the structure of the population, e.g. how many families had more than 4 children, how many persons lived alone, the age distribution of the population etc. For this purpose a punch card sorter was designed. The two most important machines for punch card processing, the tabulator and the sorter (see picture 3), thus existed, albeit in a rather primitive form, as early as 1890. The man behind this technology was Herman Hollerith (in USA punch cards are sometimes called Hollerith cards). In 1896 Hollerith founded the Tabulating Machine Company, which 1911 was amalgamated with two other companies and became IBM (International Business Machines).

(The reader who is totally uninterested in computer history and technical stuff and only wants to read about the adventures of Hålkortcentralen Siffer-Service can now jump to the paragraph just before figure 3, i.e. the paragraph that begins "But let us now...". But, I can ensure you, if you do so, you will understand less, not only about computers but of the history of Siffer-Service as well).

The punch cards began to disappear in the early 1980s and were replaced by magnetic tape (the first home computers which were introduced in the early 1980s (Sinclair, Commodore, Apple II, ABC 80/800 etc) used a standard cassette recorder for data storage while mainframe computers used much more advanced tape drives. Shortly thereafter floppy disks (which also used magnetic storage) became the standard medium for data storage (especially on home computers). They contained 720 kb (kilobyte) or 1.44 Mb (Megabyte). A 1.44 Mb floppy disk could store the same amount of data as 18,000 punch cards (a 3 m high pile of cards). And soon the mechanical hard drives, also called HDD (with fast rotating magnetic disks), became more and more common. My first Macintosh computer (I believe I bought it in 1982) only had a floppy disk drive. Pretty soon I supplemented with an external hard drive. This had a capacity of 20 Mb and the price was SEK 10,000 (in 1982 money value — in today's money value this is about SEK 27,000 which equals about 2,600 euro)! Today 20 Mb might not sound very much but one must remember that the average file size in those days was around 10-20 kb, so 20 Mb felt as a considerable amount of storage space. Furthermore the hard drives were much faster than floppy disks (regarding reading and writing data and file access time).

Another important storage medium, which was developed 1985, was the CD-ROM (stands for Compact Disc-Read Only Memory — initially they were not writeable, i.e. the user could not put data on the CD, only read the data that was stored there). These discs were used for distributing program packages, operating systems, games etc and were physically identical to CD:s intended for music. They were sometimes called optical discs, as the reading was done by a laser (no magnetism was involved here). A CD-ROM could store 650 Mb (when this technology was introduced hard disks intended for home computers could only store a fraction of that). And today SSD (Solid State Drive also called flash disc/drive — without any moving parts) is rapidly replacing the old mechanical, magnetic hard drive. They consume less energy, have a much longer life span and are lightning fast compared to their predessor (about 100 times or more faster regarding reading and writing and accessing files). USB memories are also very important and can be regarded as cousins to the SSD. The development from punch card to today's ultrafast, compact storage media has been astonishing and when I write this (october 2020) you can buy a 2 Tb HDD (2 terabytes equals 2,000 Gb) for 80 euro or less (2 Tb stored on punch cards will correspond to a card stack 4,400 km high)!

In digital computers (the standard computer today) the binary system is used to represent numbers and letters (for storage, calculations, logical operations and data management). The binary system has only two figures/numerals, 0 and 1 (which can represent either the binary numbers 0 and 1 or the logical values false and true). The punch card is an example of binary digital data storage — at a certain position on the card there is either no hole punched (0) or a hole punched (1). The basis of digital machines is switches. A switch has two positions, off and on, which can be used to either represent the two binary numbers 0 and 1 or the two logical values false and true.

0 (the binary number zero or logical false) and 1 (the binary number one or logical true) are represented in computers by different electric voltages. 0 is represented by 0 V (volt) and 1 by 5 V. This is called 5 volt logic. Voltages are never exact and for practical purposes 0 is represented by 0-1.5 V and 1 by 3.5-5 V — voltages between 1.5 and 3.5 V are forbidden as they can not be interpreted correctly by the circuits in the computer.

Initially mechanical relays were used. A mechanical relay ("mechanical" is often omitted) is an electrically controlled switch (operated by an electromagnet — click here for a short demonstration of a relay). Relay machines are relatively slow, and since relays cannot be made very small, such machines become large and draw a lot of current. The radio tube (also called vacuum tube or electron tube — an alternative terminology is radio valve etc) was invented in 1904 and was initially used in radio receivers to amplify weak radio signals. But they can also act as switches/relays and are enormously much faster than mechanical relays. So the relays in computers eventually were replaced by radio tubes. But these tubes take up a lot of space (size 5 cm more or less) and require a lot of power and for this reason emit large amounts of heat. When the transistor arrived in the late 1940s, computer development made a big leap forward. A transistor (transistor technology is mainly based on silicon) does exactly the same thing as a radio tube and transistors soon came to replace the radio tubes as they are much smaller, draw less power and have incredibly much longer life time. Plus they are lightning fast.

Mechanical relays contain moving, mechanical contacts and there is a limit how fast mechanical parts can move. Vacuum tubes and transistors have no mechanical, moving parts and are able to switch on and off billions of times every second, which is not possible with any mechanical device.

However, the really big step, the quantum leap, occurred when it was found that it was possible to integrate large quantities of transistors on one single silicon plate. This became known as integrated circuits (IC). Nowadays we often say chips (the area, south of San Francisco, that used to be the centre of chip development och manufacturing, came to be known as Silicon Valley).

The first integrated circuit was created as early as 1958 by Jack Kirby at Texas Instruments. In the year 2000 he was awarded the Nobel Price in physics for this achivement. As you don't get a Nobel Price for nothing, this hints that there were considerable difficulties to overcome during the development of these circuits. The first chips had up to 100 transistors but this number grew steadily. 1986 a chip could contain up to 1 million transistors (or more correctly transistor functions). And this progress has gone on and on.

Today you can fit billions of transistor functions on a few square centimeters. Which means that, for example, a modern smartphone has immensely much greater computing capacity than a mainframe computer in the 1960s.

The chip that had most transistor functions in 2019 was a 1 Tb eUFS RAM-memory from Samsung with 2 trillions (2 million million) transistor functions! Each transistor has exactly the same function as a radio tube. In the beginning of this article there were some quotations with early predictions regarding the future development of computers. In one of the quotations was mentioned a computer with 18,000 radio tubes (equivalent to 18,000 transistor functions). The weight of this computer was 30 tons. A computer with 2 trillion radio tubes should have a weight of over 3 billion tons (if we make a simple proportional calculation)! Furthermore, such a computer would never work, as thousands of tubes would always be non-functional because of the short life span of radio tubes (about 500 hours). And if you replace those broken tubes and connect the computer again, thousands of other valves would already have stopped working.

But let us now, after this summary of computer development (as you can see I do not save any effort when it comes to educating my readers), return to our story. The story about my uncle and his computer company. The computers used during WW2 for code cracking operated with combinations of relays and radio tubes for their calculations. For administrative data processing mechanical machines with gears (cogwheels) were used initially and later machines with combinations of gears and relays. The first machines that Hålkortscentralen Siffer-Service bought came from the above-mentioned French company Bull. Which was perhaps to expect, since Erling had worked with Bull machines at the Swedish Pensions Agency and was used to them.

Figure 3. Siffer-Services' first machine park, Bull Series 150. The first version of this series was introduced as early as 1941 and was completely mechanical. Later versions (i.e. in all probability the one that Siffer-Service bought in 1954) used small relays. To the left we see the tabulator, which was the computer unit itself, where the calculations took place. In the middle of the top of the tabulator we see the printer and just to the right of this the punch card reader/puncher. I remember that the cards were read at a furious pace (about 3 cards per second). The connection panel (Kopplingspanel), indicated by an arrow, will be desribed more in detail below (see picture 4).
The machine on the far right is a card sorting machine, which sorted the punch cards according to the criteria set. This was a very important machine in the "punch card world". The sorting was done after the columns on the cards. One column at a time. A complete sorting, after all the 80 columns, required that all current cards were run 80 times through the machine (the sorting machines were running almost continously). Click here for a short demonstration of a card sorting machine (15 seconds).

Siffer-Service quickly gained many customers (for several years they were almost alone in their niche). I remember that one of the biggest was Metall (one of the big unions in Sweden — at that time they had a huge number of members as the steel and metal industry in Sweden was a dominant part of the economy). In general, my uncle preferred many small customers to a few large ones. In the latter case, the business is very much affected if a customer, for example, goes bankrupt or goes to a competitor. If you have many small customers instead, it is hardly noticeable if one of them disappears.

Some curious: Erling once told me that the first tabulator he bought from Bull had certain limitations. I don't remember exactly what the problem was. It might have been something about that there were some types of divisions that were not possible to do. Erling had found a smart solution to this problem (if I remember right the solution was purely mechanical). When Monsieur et Madame Paul from Bull once visitied Siffer-Service, Erling showed them his solution and Monsieur Paul exlaimed, "Mais monsieur Kaiser, ce n'est pas possible!" ("But Mr Kaiser that's not possible").
As the tabulator partly was mechanical it was possible for even a non-expert to find solutions to certain problems (as Erling did). But that time is by far long gone. Which is to be expected, but it is still a little bit sad. Cogwheels and relays can be seen by the naked eye and a smart person can understand how everything works, without being an expert. The function of transistors and chips is impossible for the eye to see and the function of these components is very, very complicated. If you really want to know how a transistor works, you must have a Ph D in solid state physics (I am not joking).

A reader, interested in computers, may by now wonder how these early machines were programmed. Well, this was done through large plugwires panels with several thousand holes (terminal holes), which were connected by a complicated network of wires (cables).

Figure 4. Close-up of a plugwires panel (the connection panel). The cables represent the program flow. This "programming language" could be called "Cabol" or "Cobel" ("Cobol via Cable" — for any reader who lacks the sense of humor, I can tell you that I'm joking). A wire/cable can represent a goto statement in a program (e.g. goto line 232"). The plugwires panel with its cables thus represents a program that performs a certain task, for example calculating salaries and taxes etc for a certain company. Figure 3 shows the location of the plugwires panel (Kopplingspanel) on the tabulator.

My uncle was apparently extremely good at this type of programming and he did all the programming himself. At the time they used a sheet of paper (about 420x 594 mm), where all the holes on the plugwires panel were printed (the paper was an exact copy of the plugwires panel) and the person who was to make a program drew the connections between the different holes with coloured pencils. As can be seen from the picture above, these connections could be quite complicated. Then there were employees, whose task was to connect the cables based on the master on the sheet of paper. A job that required accuracy and concentration. A single connection error and the program would not do what it was supposed to do. I remember the panels as being rather heavy (I was 10-12 years old then so an adult might not have thought they were heavy). The width of the panel was maybe around 35-40 cm and the height around 60 cm. Some programs were run only once, which meant that someone afterwards had to remove all the cables and sort them into different boxes by length and colour (it was my job sometimes — so I can brag that I worked with computers already in the 1950s, which most people do not believe (yes, it is a slight exagerration but not entirely a lie). I got, by the way, into an argument with a Portuguese recently because he was sure there were no computers in the 1950s. But he was evidently wrong as Siffer-Service was founded in 1954 (see picture 1 — open the enlarged version of the picture).

At Siffer-Service there were lots of plugwires panels and for jobs that were run regularly (salaries for different companies were of course run once a month) the plugwires panel was saved with all its cables in a storage room. When it was time to run the job again, someone collected the current plugwires panel and then attached it to the panel holder on the tabulator (see picture 3 above — the arrow points to where the plugwires panel was located). And then it was just to run the program. The data was on punch cards. For every person, whose salary was to be calculated (if we take salary calculation as an example), there was a punch card with contracted salary, overtime, days off etc, etc (all information that was needed for the calculation to be made). These cards were read by the tabulator and the result was printed by the printer on the tabulator. New punch cards were also punched by the tabulator to be used the next month and for balancing the books etc.

Eventually more modern machines came and I remember when uncle Erling got the first Electron Brain (as it was called) and proudly demonstrated it to me (by this time I was probably around 20). The official name was Bull Gamma 3. This dramatically speeded up the calculations and was the dawn of the modern computer. Now new programming techniques began to appear, using punch cards instead of plugwires panels, and my uncle now chose to hire university-educated programmers to deal with that side of the business, so that he could concentrate on the company's development. And besides, by this time Siffer-Service had grown so much that several persons were needed just to do the programming.

Figure 5. In the background we see "manager" Erling Kaiser himself (he looks very young to be around 40 — I believe the picture is retouched, this was very common in those days). The picture shows Bull's Gamma 3, which was Bull's first calculator with radio tubes. The Gamma 3 was a complement to the standard punch card machines and was connected to the above-mentioned tabulator (Series 150), shown in figure 3. Gamma 3 came as early as 1953, but it probably took a number of years before Siffer-Service could afford to buy one. On top of the Gamma 3 is an oscilloscope, which gives the right impression of advanced electronics and mystic science (in all old James Bond movies oscilloscopes always are abundant in the lair of the supervillain). Click here for a brief description of Gamma 3 (less than a page).
The machine in the foreground, with the two thick cables, was handling punch cards and was called a PRD (Punching, Reproducing, Duplicating).

Erling was very careful to emphasize that he was not a CEO but a manager. He meant that a CEO executes the board's decisions (i.e. is the board's "slave"), while a manager manages his company. He owns it. And does what he wants with it.

The word "manager" should be taken figuratively here. My uncle used the Swedish word "disponent", which I have translated to "manager". As business structures in those days were very different in Sweden compared to England/US, it is difficult to make an unobjectionable translation. Erling was playing on words here, which is often impossible to translate. But I think the reader, from the context, will understand what Erling tried to say.

When Erling was young he worked for a time at a small dairy in Denmark. The owner called himself "mejeriejere" ("dairy owner"). It was evident that Erling liked this, as he often talked about this dairy owner. He regarded himself as the owner of Siffer-Service (which he of course was, but today you are not supposed to use such vocabulary). It was his company. End of discussion. Erling belonged to the old school and would probably have had a hard time adapting to today's business climate, where the right values and giving the employees courses in sexual and other minorities rights etc are more important than having good products and running the company profitably.

From the (political) left it is often maintained that the employer certainly has invested his capital but that the worker (employee) is investing his working hours in the company. Thus the worker is as important as the employer for the company and these two should split the profit equally. Erling thought this wording was deceptive and used to reformulate it as "The employee invests his working hours but the employer invests his capital plus his working hours. Regarding rather small companies the employer is often forced to take large loans which he is personally obliged to repay. That means he takes a great risk (if the company fails he might lose everything; house, car, summer house and still have large sums to pay back). And someone who takes great risks deserves to be rewarded if he is successful. Furthermore employers often work much longer hours than the employees. I remember that Erling often came home from work very late and had headache and was very, very tired. And he spent many Saturdays or Sundays at Siffer-Service. Erling never denied that the employees were important. But, being an employer (particularly in a small company) is tough. Much tougher than many employees might believe. And if you are not satisfied with your share as employee, you can always start your own business! And besides, if the employees want to share the profit, they should also feel obliged to share the companys possible losses and debts. Right? Which I suspect few employees are prepared to do.

Anyhow the development of computers continued and eventually Siffer-Service switched to machines from Univac. Now we are talking about computers in the modern sense, which were programmed with modern programming languages similar to what we have today (one of them was Cobol, which for a long time was the most important programming language in business, finance and administrative computer applications and still is in use today, 60 years later — impressive right?).

Figure 6. Meeting at Hålkortscentralen Siffer-Service. The picture was taken in 1962 when Siffer-Service signed an important contract with Siemens. Which one is uncle Erling can hardly be missed (he is the hub that all spokes point to, i.e. the seated man with pen in hand). On the table is a tin box with Wilhelm II cigars (the white arrow), which was the brand Erling always smoked. At this time, restaurants, meeting rooms and coffee rooms were veritable gas chambers because of all the smokers.

In 1974, when my uncle Erling was 55, he sold his company for a number of millions and retired.

My cousin Birgitta told me that one of the reasons my uncle sold the company was that the premises on Kungsgatan had started to get too cramped. Siffer-Service had to grow in order to keep up with the competition. And Erling did not want to move to any industrial suburb. He loved Stockholm's inner city (especially Norrmalm) and the environment there. He wanted to have his lunch at Hotel Carlton and do what he was used to do. And liked to do. So he left the move to those who were to buy the company (after the sale, Siffer-Service moved to Solna, a suburb north of Stockholm).

He lived until his death in a two floor penthouse apartment out in Hagsätra (south of Stockholm) with a wonderful view and enjoyed life. I visited him sometimes and then we went to "Stan" (Stockholms Centrum) and had lunch (often at Ulriksdals Värdshus) and then we went back to Hagsätra and smoked cigars and drank brandy and played classical music. And talked about life and many other things. Erling had some very interesting stories to tell. Since he was Danish, he could only travel to German-occupied territory during WW2. Among other things, he worked for a year in the town of Gleiwitz (Gliwice in Polish) in southern Poland (about 50 km west of Auschwitz), driving trams. In Gleiwitz there were, by the way, three satellite camps to Auschwitz.

Let me just give a few examples of Erling's memories from Poland. He was a very punctual man and when he drove the tram he always tried to keep to the timetable. At one point, as he was about to leave the terminal station, a senior SS officer came running and signaled with his hand, from quite a distance, that the tram should wait for him. Which Erling did not. He drove. Punctuality is a virtue! A few stops later a Wehrmacht (the regular German army) officer left the tram. On these trams you entered in the back (where the conductor was) and left at the front door. As the Wehrmacht officer passed Erling, he said in a low voice, "Seien Sie vorsichtig, Herr Kaiser!!!" (I guess Erling had a nameplate on his chest), which means "Be careful Mr. Kaiser !!!". Now I do not think there was any real danger, because punctuality was a virtue for the Germans. If an SS officer came too late, he had to blame himself and I do not think Erling would had been punished if the SS officer had chosen to report him. But who knows, the SS had great power in Poland. Had Erling been a Pole, it might have been very dangerous to do what he did, but he was Danish, i.e. Aryan. An SS officer could easily shoot dead a Pole on the street if the Pole obstructed, but if he did so with a Dane, he would probably have been given a severe punishment, and had perhaps even been executed.
During his time in Gleiwitz my uncle became friend with a young Pole and they used to listen to the BBC shortwave radio transmissions. Poles who were caught listening to foreign radio stations were sent to Auschwitz I, where they were gassed or had to work themselves to death. What would have happened with Erling if he they hade been caught red-handed is an open question. Maybe he would have been sent back to Denmark. Or maybe something worse would have happened. On several occasions BBC reported that Jews were being systematically killed by the Nazis. He also told me that he had heard from people living close to Auschwitz, that they could feel the smell of burning flesh all the time and that rumours circulated about terrible things happening in the camps. My uncle never denied the Holocaust. He had been neighbour to it during about a year so he knew it was a reality.
In Gleiwitz, people burned coal for warming the houses and the air during the winter was so bad that the locals, according to Erling, used to say, "Inhale three times and spit out a briquette".

Erling used to say that if he were to write his autobiography, it would have the title Pure Luck. He believed that he had been very lucky. E.g. that he started working at the Swedish Pensions Agency just when they were going to computerize and in this way, without any merit of his own, became expert in computers. In additiion to that, through his father in law, he got the opportunity to save the shoe shop and because of this was able to borrow money from the bank (we are talking about large sums). A young, unknown Dane, without any formal education (only a junior high school diploma corresponding to year 9 today), would never have been able to borrow so much money otherwise. Erling's strength was that he took advantage of the opportunities that arose.

Different school systems in different countries. Exams, grades etc are often very difficult to translate from one country to another. Fundamentally what I am trying to say is that Erling went 9 years to school (which means that he was about 15 when he finished school) and had no college or highschool or university exam or degree.

Bill Gates has said much the same thing. Had he been born a couple of years earlier, it would not have been possible to develop DOS, because then there would have been no hardware to run DOS (DOS stands for Disc Operating System and was Microsofts's original operating system and the product that laid the foundation for Microsoft's success). And had he been born a couple of years later, then DOS or something similar would already have existed (developed by someone else). And this probably applies in many cases. Those who succeed may be talented and skilled and diligent, but without a good portion of luck, they may not have achieved the enormous success they did. And as said above, their strength is that they take advantage of the opportunities that are offered them.

All in all, Erling Kaiser was undoubtedly a computer pioneer in Sweden, who probably founded the first privately owned computer company in our country. Most of the CEO:s today, who come from Handelshögskolan in Stockholm (the most prestigious eduction in economy offered in Sweden) go straight to a "table ready laid" (an already existing company that has attractive products and good economy) and immediately get grotesque fantasy salaries although they have never created anything or invented anything or taken any risks, Unlike them Erling started with two empty hands. He built his company by himself. From the very foundation. He started from zero, Talent, a strong will, hard work and a portion of luck were the keys to his success! Well done uncle Erling!!!

This was a small exposition over the infancy of the computer world, from a Swedish perspective. I hope you found it interesting! If nothing else, I thought it was a little fun to refresh old memories from the past. Some of what I have written above may not be completely accurate (memories fade with increasing age), but I think most I have written is correct enough to give the reader a fairly good picture of the magical years when the modern computer grew to maturity.


Back to Krister's english Internet page
© Krister Renard