Alan Kay: “Could the ancient Romans build a computer?”
  • Transfer
This is a very interesting question - similar to the one I used to ask in a computer design class: “What kind of computer would Captain Nemo develop for Nautilus? “Draw its design!”

Today’s question can be understood differently: for example, in the sense of “With the technologies available at that time, could a complete genius, worse than Leonardo, think of a device that we would call a digital computer?”


Hublot's miniature replica of the Antikythera mechanism

This approach also allows the use of "Greek thought", since the Romans had many Greeks - free and slaves - who were engaged in deeper thinking for these people with practical thinking. This is a good combination, because the Greeks had little interest in technology, and the Romans were interested in - some of the most interesting mechanical inventions of the Greeks were toys, especially those created by Alexandrian Greeks.

Let us not take into account the fact that one of the problems of slavery is that it does not motivate people to invent mechanisms for performing a large amount of elementary work, including calculations. And let's not take into account the question of the “accumulation” of the ideas of predecessors that are needed and used even by geniuses - the essence of the question actually sounds like, “Can we build something?” At this point in our history.

The translation was made with the support of the company EDISON Software, which is professionally engaged in the development and testing of software .

Somewhere else in other answers it was correctly written that both the Greeks and the Romans carried out calculations using a kind of abacus (the word “calculus” is the name of the stones moved through their counting tables). It is important to note that the abacus really had the meaning of "zero" embedded in the way it works.


A reproduction of the Roman “pocket” abacus that would fit in the pocket of a modern shirt. The "stones" here are not stones, but beads in sliding slots.

Naturally, I, an IT person, was very interested in Babbage (and Ada), and was initially very impressed with the inventions that he had to create just like his predecessors (for example, a careful mechanical drawing, a turret lathe, etc.) .) as well as inspiration from a jacquard loom. It was fun to launch a reproduction of the Difference Machine in the London Science Museum (an ultra-precise design, but with a lot of negative consequences and efforts).


A few years later I had the opportunity to truly have a look at an early Jacquard loom, and I was completely amazed at the amazingly excellent approach to mechanical design; even the organs of that time with huge mechanical pipes were not so wonderfully made!

I realized that Babbage chose a very bad engineering approach — he was a mathematician and tried to do things “for sure” —but in reality he would have to look at the actual mechanisms of the jacquard machine much more closely.


Only the top 1/3 is a jacquard mechanism. The loom itself is mainly made of wood with a small mechanism (it used the logic of "hangers" / "coat hanger").

Jacquard went completely in the opposite direction: everything just falls apart, unless he needs precision (most of the accuracy he received was in the “guide holes” that limited the dangling parts, but

allowed them to dangle on both sides (one of the assumptions: had to periodically “reduce errors” in the chains of causes and effects).

As a result, the distribution and increase in the number of errors is practically absent (the Babbage scheme causes a huge distribution of errors). One of the many fun moments of the Jacquard approach is that this mechanism does not require the invention of a new type of looms: the jacquard machine is on top of the existing loom and simply pulls the threads in accordance with the program in the chain of punched cards.

Thus, one person operating a jacquard loom can work on thousands of threads with punch card programs using only human power.

It is interesting that there is nothing mechanically complicated in the jacquard machine - most of the logic can be built from hangers! - and this could be done using technology alone, available on both sides from 0 BC. Here is a brief description of some of the technologies available .

The Wikipedia article on the jacquard loom is the beginning, but not enough details. I'm still looking for a better description of the jacquard mechanism online (please help!). Here the actual details of how Jacquard saw that the choice does not require careful accuracy are important, unless it is really necessary. In the meantime, read this pdf starting on page 5. And watch this video on YouTube to see how different parts operated by one person work.

There are several books containing useful details: Essinger, James (2004). Jacquard web. Oxford University Press, Oxford, and especially: Bradbury, Fred (1912). Jacquard mechanisms and harness installation. John Heywood Ltd., Technical Book Depot, Halifax, Yorks. The best book I found was: (1888) "Jacquard Machine, Analyzed and Explained", E.A. Posselt. Thanks to the wonderful Brewster Kale and his, this book can be found and downloaded via Google.

I think it would be quite possible to use modern thinking to design a programmable computer based on jacquard mechanisms and then assemble it using only methods, tools and materials available until about 0 BC.

Next week, I will try to find a clear online report on how the jacquard loom actually works (it almost always happens that museums — even “science and technology” museums — just expose artifacts and don't give explanations or demonstrations * like * and * why * they work).

Appendix I

I forgot what I wrote about Babbage for Quora, and mentioned there more details about the jacquard loom. What is the best technical book about Charles Babbage ?

I found on YouTube a pretty good video from the V & A Museum, which has a three-dimensional animated explanation of Jaccard's invention.

It seems that he does not have a sound track, but I think that it can be understood only by visual explanation. You may want to watch the animation in half the real speed of the video series, in order to have time to understand and follow the causal link.

Appendix II

Marcel Levy asked the question that should be put here in the main text: “That is, you say that the problem was not in the practical side, but in the theoretical side?”

I think so. It is worth looking at some of the mechanical inventions of the Alexandrian Greeks, including: an organ with labial tubes, which they played on using keys with wind, stabilized by water pressure, etc. Also Greek theater mechanics, etc. And, of course, the astronomical calculator Antikythera.


The Hero of Alexandria was the most interesting inventor of mechanisms around 0 BC, besides the organ with labial tubes, an odometer was reported that counted in a digital format, and, most interestingly, a programmable trolley (recreated here on YouTube).

What else did Geron do on the model of these “toys for amusement and surprise”? He had the idea of ​​"programming", and probably he developed his other forms for automation, which he invented.

Looms have existed for thousands of years, and the most complex of them - hundreds.

The Greeks (and then the Romans) understood that the cultural “official” way of writing numbers was “inconvenient”: we can see their real thinking by looking at their computers (various abacus).

We could definitely ask, “what do they really want to figure out what goes beyond their abacus?”. At first, Babbage was guided by inaccuracies in the mathematical tables (“I pray to God that these calculations should be made by steam!”). Turing was initially guided by the results of Gödel. The study of methods of physical computation has been driven by the needs of the war since the Second World War.

They definitely had brains and much of the mathematical and physical outlook. It seems to me that most of all they lacked a sense of need that would force them to use what they knew to see how to make the machine perform calculations for them.

Also popular now: