23 thoughts on “Ancient Roman Computers”

  1. The book Unknown Quantity – A Real and Imaginary History of Algebra has a fair number of examples of ancient mathematical symbology. Hardly shocking that people stayed mostly focused on geometry and ‘logic’ remained linguistic logic through that era IMO.

    1. That is a fantastic book, as much as I disliked John Derbyshire’s ranting on the Corner, that book was brilliant.

  2. The Romans may not have had a zero, but they had an abacus. It’s doesn’t seem too big of a jump to get to a zero from an abacus.

    Comments in the article raise questions about whether the Baghdad Battery was even a battery. Without electricity they’re probably SOL (another comment talked about fluidics, which is kind of an interesting approach, but I doubt would be any more practical).

    I suppose one question is just how simple a structure you can build and still call it a computer. An XOR gate is kind of like a computer–you could call one input data and the other input the program (the program is relatively simple–invert or don’t invert the data). It’s binary, has separate program and data. Not particularly useful. No memory.

    Anyway, I suppose the question is whether a modern Martin Padway or Connecticut Yankee could make a practical computer if they found themselves in Roman times. It’s kind of an interesting thought experiment. We’d have to postulate someone a bit more technical than Martin Padway–he couldn’t make a working clock, despite having a working mechanical watch to crib from.

  3. My first thought was the tinker toy computer. However,

    the first problem is understanding what a computer is. We’ve ruled out analog which simplifies things. It’s not about electrical or mechanical components either. It’s about implementing the logic of a cpu.

    Roman numerals are irrelevant if you give them the required logic which we may describe as ones and zeros but really has nothing to do with ones and zeros. It’s simply two states. State * and state not *. Even combining them into places is not strictly about numbers but a counting of states. It’s simply more convenient to think of them as binary digits or bits rather than binary states. Two bits gives you four states; three gives you eight and so forth.

    So, the Romans could start with a memory store which can be anything their technology allows them to build. It simply has to allow changing of states in a sequential block of memory arranged in units of whatever size they choose.

    Then they need to assign logical actions to various states.

    Doing this, they could make a computer out of a lot of rocks and pits with slaves running the length and following the rules. They don’t even need a drummer.

      1. Ken,

        Exactly. The Romans never followed up on steam power, hydraulics, computers or over technology because they had an endless supply of slaves. Another reason was that the use of those technologies was limited to the priests who used them to “wow” the masses with their magical abilities…

          1. Ken,

            And advancement of technology is pulled by economics and markets, not pushed by research or invention.

            Coal for example was known to the Romans, but they saw little use for it as a heat source, so it wasn’t until the wood shortage in England in the 1600’s triggered a demand as an “inferior” replacement for wood that the Industrial Revolution started.

  4. There displays would be huge, slow and only be seen from a distance. It would have slaves running from the rock pits to the coliseum with three rocks painted red, green, and blue.

  5. Turned off the computer. Made myself lunch. Then struck by a thought. Perhaps this explains pictograph written languages!

    We don’t need Romans. Cavemen could produce the computer I describe and with a pictograph language it could be a very powerful computer even at slow clock rates.

    That’s because the instruction set does not have to be limited to what a common cpu includes. The instructions could be anything (Kill the king!) These instructions could be carved in stone as pictographs. The instruction pointer (slave, devoted servant of the computer god) refers to the pictograph carved in clay associated with the instruction state of the current pebble pits. The memory locations do not even need to be the same width. Whenever the programmer needs an instruction he just adds that state to the carvings in the book of clay. If they start out with a three bit computer (8 states), but need a 9th action they simply add a pit just to the current memory location. They never run out of memory either… it’s just more consecutive pits. When they need a 17th action, that memory location gets five pits and they now have 32 states they can carve into the clay.

    By the time you get to the Romans, they would be so far advanced beyond us (they use more primitive computers to design newer technologies.)

    It’s all about the free mind.

  6. Turned off the computer. Walked away vowing not to return for hours. Met a roman tech guru that had an innovation to increase the refresh rate of the coliseum screen by…

    Screen memory never needs more than two pits or four states per pixel (R,G,B,none) Several slaves are assigned to each pixel. All they do is continuously monitor that the pixel on the monitor matches the pixel state for the particular memory location (they really don’t even need to do that, they simply run the current state to the monitor even if it’s the same. Your refresh rate is determined by how many slaves per pixel you have. Each pixel location is handle in parallel independent of the system clock rate.

  7. Well there was that Antikythera Mechanism that found in a Greek shipwreck. An attempt to reconstructed one based on 3d scanning of the original believes it was used to calculate the positions of the planets and the timing of eclipses.

    1. That’s not a very practical writing system–24 possible colors, 2 types of materials, 2 directions of twist. That’s 96 types of yarn that a khipu writer must have had on hand. There are supposedly five other degrees of freedom available to the khipu-writer (handedness of the knot, length of the string, number of pendants, some others which the article doesn’t call out). I suppose the telegraph office could keep all the different types of yarn on hand for composing a message, but it doesn’t seem at all useful for everyday writing.

      I’m skeptical that some of the items he thinks are distinctive really are, but maybe he’ll break the code sometime soon.

      The article calls the khipu a type of abacus, which seems to be a fundamental misunderstanding of the role of an abacus. An abacus is dynamic, and used to carry out calculations. I don’t see how a khipu can help calculate–it looks more designed to store information.

      1. Pictographic languages aren’t very practical either, but they exist. I think you are exactly right that it is for storing information more than as a general means of transmitting it.

        He’s saying it’s a written form so when he mentions abacus it’s not meant to mean a calculating device, but that the information recorded is done in a similar way. The article says they’ve already determined that the knots contain semantic information. The way the strings are organized reminds me of the pages of a book (where a single strand would be more like a scroll.)

        The striking thing is the density of information; of course, not all combinations may contain info any more than a random combination of letters would.

        If you need to record a lot of information but don’t have paper or printer this might be a good solution. Imagine if we could read it?

  8. BTW, if you think pits, rocks and slaves (PRS) would be meaningless as a computer, you’d be wrong. I remember cpu’s in the 70s working in kilohertz. That’s within the range of a parallel slave computers potential. All it takes is for elements of the PRS computer to be upgraded over time with mechanical devices to replace the labor. Now you have a better device to design the next generation. If they figure out how to go from mechanical to electronic things begin to move even faster (they may do fluidics as an intermediate step.)

    Again, it’s all about the mind contemplating the possibilities. A good bit of the scientific method and experimentation doesn’t hurt either. If you tried to explain how a computer works to most people that hadn’t seen one, even today, you get blank looks. That doesn’t make the computer any less powerful and useful.

Comments are closed.