The foundation stone for the computer was laid 200 years ago

Electrical engineer Prof Dr Dietmar Tutsch remembers Charles Babbage, whose pioneering idea of an analytical machine was forgotten

Today, it is impossible to imagine society without computers. But the groundwork for it actually began 200 years ago. Charles Babbage, mathematician, philosopher and inventor, created the first calculation concept for a computer in the early 1820s. Electrical engineer Dietmar Tutsch has studied Babbage, whose ground-breaking invention was nevertheless forgotten.
"Charles Babbage was an English scientist and a multi-talent," says Tutsch, "he became known as a pioneer because he created the template for designing a computer." Although the machine was never realised, he was awarded the gold medal of the Royal Astronomical Society for this calculation concept exactly 200 years ago, in 1824.

Babbage developed the difference machine in 1822

As a precursor to the analytical engine, Babbage first developed the so-called difference engine in 1822. "This machine was used to analyse mathematical polynomial functions, i.e. more precisely, to calculate their function values. Polynomial functions are used for simplification purposes to approximate or replace more complex functions. These functions are then often presented as tables in mathematical reference works so that they can be read off directly," explains the expert. Babbage therefore built a machine that calculated precisely these tables and, of course, functioned purely mechanically at the time.

The concept of the analytical machine

This was followed in 1837 by the design of the Analytical Engine. Tutsch explains: "In principle, this was a further development of this difference machine into a mechanical calculating machine for more general calculations using algorithms. That was absolutely new! Until then, a step-by-step procedure had always been used, i.e. one calculation after the other, without any influence or decisions being made during the calculation." Tutsch knows that automatic loom control systems, such as those used in the textile industry in Wuppertal, also functioned according to this principle. "The new thing was that there were leaps in the design. For example, depending on a previous result, calculations could be skipped, i.e. omitted, or you could jump back to a previous calculation point to perhaps do the same thing several times but with different numerical values. These jumps in various forms are the core of every computer programme today." As the machine was never built, he says, one can only speculate about its appearance, although one must assume that there are many precisely ground individual parts, such as gears and spindles.

Ada Lovelace - the world's first female programmer

Ada Lovelace, daughter of the romantic poet Lord Byron, was partly responsible for the later success of his calculation concept. Known today as the world's first female programmer, she contributed the precise functional description of the machine. "Ada Lovelace recognised the added value of this analytical machine, these algorithms, which made it possible to do much more than just calculations. She drafted a precise description of the analytical engine, including a programme algorithm for calculating Bernoulli numbers (Bernoulli numbers are among the most important constants in mathematics, editor's note)." Tutsch emphasises that the fact that she was able to make and realise this invention as a woman at the time is particularly noteworthy. The programming language Ada, which was used in real-time systems, i.e. computer systems in which temporal conditions had to be fulfilled, was named after her.

Doubts from scientific experts prevented funding

On the recommendation of the British Association for the Advancement of Science, the analytical machine was ultimately not built. As the machine would have consisted of many individually manufactured parts, experts doubted its accuracy at the time. "This made it impossible to estimate the costs," says Tutsch, "so they advised against building it and did not provide any funding."

Babbage the inventor

You might think that Babbage would have retired in frustration, but this was not the case, as he had done crucial groundwork in various other areas. We now know that he was the first scientist to decipher the Vigenére cipher, an encryption method for text messages dating back to the 16th century, which he did not publish throughout his life. "And there are a few other things to mention," says Tutsch. "He travelled a lot in the railway sector and investigated track gauges. He invented the so-called cowcatcher on the front of locomotives or the test bench wagon for locomotives, which is used to measure the performance of locomotives. In medicine, he invented the ophthalmoscope, parallel to Hermann von Helmholtz, whose version is mostly used today. He was also very busy with statistics and tables of everyday events."

Science also forgets

The overall concept of the analytical machine was forgotten, confirms Tutsch, and even computer pioneers such as Zuse, Aiken, Eckert and Mauchly only came across Babbage after their inventions. "We can only speculate about the cause," he says. "Back then, researching the literature was quite tedious. There was no internet or search engines. It was all done via special books with index directories in which you had to search for the right keywords. It was easy to overlook something or simply use the wrong keyword. That's why the publications of Charles Babbage or Ada Lovelace simply couldn't be found."

Adequate calculating machines did not achieve the accuracy of the analytical engine until 1960

Around 140 years after Babbage's calculation concept, calculating machines achieved the accuracy of the analytical machine, but the actual beginning of the computer age was not so much to do with the accuracy of the calculation, the scientist explains, but more with the reliability and cost of these machines. At this point, the transition from mechanical to electrical computer components and the invention of the transistor were decisive. "Mechanical manufacturing tolerances and problems led to a rethink towards the electrical realisation of a computer in the 1940s. However, the electron tubes with which these were constructed failed too quickly and too often." Tutsch cites a famous example: "The ENIAC (first electronic universal computer, editor's note) from 1945 had 18,000 electron tubes. With an average lifespan of 1000 hours, one of the tubes failed every 3.3 minutes. So you could calculate for an average of 3.3 minutes. And then you had to find the failed tubes in the house-sized computer to replace them. The ENIAC belonged to the so-called first generation of computers. The second generation was created around 1955 with transistors, the third generation around 1965 with integrated circuits (ICs), where you could fit a lot of transistors into one component, and the fourth generation in the early 1980s with microprocessors made from highly integrated circuits."

From Babbage's pioneering work into the digital age

The algorithms that were recognised 200 years ago, with which a machine can do much more than just perform calculations, are now an integral part of technical processes. Dietmar Tutsch heads the Chair ofAutomation Technology / Computer Science in the Faculty of Electrical Engineering, Information Technology and Media Technologyat the University of Wuppertaland says: "At the Chair, we use computers to automate production processes in industry on the one hand and to realise so-called embedded systems on the other. These are technical products that are controlled by a small computer that is usually invisible to the end user, e.g. a washing machine, a fully automatic coffee machine, a heating system or many things in a car, such as the airbags or the ABS. The algorithmic processing of the data is central to this." If, for example, the sensor on the front bumper of the car reports a collision, the airbag is automatically activated. And this happens at exactly the right time, so that it is fully deployed just before the head hits the steering wheel. "Even the data-driven processes of AI require algorithmic processing, e.g. to describe and simulate the structure and learning of a neural network."

Since 2010, there have actually been plans to build the Babbages machine after all. "Yes," concludes Tutsch, "because there are still many uncertainties. We want to fully understand the Babbages machine. Many details are still unclear or controversial." In the 21st century, scientists are constantly turning to historical models in order to use their findings to shape the future.

Uwe Blass

Prof Dr Dietmar Tutsch is Head of the Chair of Automation Technology / Computer Science in the Faculty of Electrical Engineering, Information Technology and Media Technology at the University of Wuppertal.