▲ 39 points
back
22 comments
So, this is a special purpose measurement tool with some automation, and it is not a computer. The way the word computer is understood today is an automatic, general purpose, stored program, electronic computer. The Antikythera mechanism, OTOH, is a complex, slightly automated, astrolabe. There is a very fundamental conceptual leap that the the astrolabe fails to make which is that all calculation can be done by a machine made of rather simple parts: switches.
All of that said, I think the main reason that the Antikythera mechanism continues to be fascinating is that people have a very haughty view of the modern world and modern civilization, and they assume that people of the past were dullards (with zero evidence to support the view). I would wager that people have always been rather smart, but that we’ve merely become increasingly specialized over time.
> this is a special purpose measurement tool with some automation, and it is not a computer... The Antikythera mechanism, OTOH, is a complex, slightly automated, astrolabe.
This is entirely wrong, and I don't know from where these ideas come. Nowhere in the article are the words "measurement" or "tool" used. The Antikythera mechanism does not take measurements.
>> This little instrument not only predicted eclipses, but also calculated the phases of the moon, the movements of prominent stars and the revolutions of the planets – although not with accuracy.
If the Antikythera mechanism does this, then it most definitely is a computer.
It's a tool of measurement in the same manner in which an environmental simulation is. A reproduction of a thing for the purpose of study is a form of measurement enhancement. This is also a tool in the same manner a hammer or a computer is a tool.
Special purpose computational tool, perhaps. A computer in the modern sense of the word, absolutely not. It is neither general purpose nor is it Turing complete.
Please note that my initial reply stressed the point of a modern understanding and use of the word computer. If people wish to call this a mechanical computer it is more similar to a differential analyzer and yet still less capable. As human beings alive after the Second World War, we have far more terms with which to accurately categorize mathematical tools. Calling something more similar to an astrolabe a computer is akin to considering a hand saw and a circular saw the same thing.
> It's a tool of measurement in the same manner in which an environmental simulation is.
I don't see how either is a "tool of measurement." A ruler, scale, thermometer... these are tools for measurement. A simulation doesn't measure anything. The purpose of the mechanism is not to measure, it is to predict. It calculates, and that makes it a computer.
> A reproduction of a thing for the purpose of study is a form of measurement enhancement.
The issue is not with the use of tool, but I don't see how measurement enters into it. Measurement enhancement is not a general form of anything. A simulation is not for measuring, but for evaluating the effect of changes, for evaluating models, for tuning performance or safety, or for testing. Though measurements may be taken within the simulation, the simulation itself does not measure anything.
> Special purpose computational tool, perhaps. A computer in the modern sense of the word, absolutely not. It is neither general purpose nor is it Turing complete.
There is not a modern definition so much as there is context. You're defining computer to mean a programmable electronic device, I think you mean Turing machine, but this is a very specific meaning for a very general term. A computer is simply anyone or anything that computes. In fact, a computer can be anything from an abacus to a slide rule to a calculator (analog or digital) to even an automatic transmission. "Computer" does not necessarily mean anything electronic unless context dictates that it does.
> Please note that my initial reply stressed the point of a modern understanding and use of the word computer.
There are plenty of computers still waking around today that are not digital, and many of them we call computer scientists.
> If people wish to call this a mechanical computer it is more similar to a differential analyzer and yet still less capable. As human beings alive after the Second World War, we have far more terms with which to accurately categorize mathematical tools.
An astrolabe is not a computer, it is an instrument for making astronomical measurements. Instead, a computer is anything (or anyone) that computes, calculates, or reckons. We can be more specific by qualifying analog computer, or mechanical computer, or microcomputer, or even an accountant, but they are all computers. The category error is not necessary; only context matters here.
> Calling something more similar to an astrolabe a computer is akin to considering a hand saw and a circular saw the same thing.
That they are both tools does not mean that a screwdriver is also a wrench. They are both within the category of tools, but not all within the same category are of the same type.
Here is another category error:
Socrates is a man.
I am a man.
Therefore, I am Socrates.
This is an unhelpfully reductive definition of computer. The name computer used to mean people who compute things, but it’s also entirely reasonable to call a device which performs computations a “computer”, including tools like an astrolabe, slide rule, or nomogram.
Regardless of where this thread ends up on the definition of 'computer', I guess the more interesting question is: how close might these folks have been to building something like Babbage's engine?
If you had to introduce programmability, memory etc. -- no matter in how rudimentary a way -- how (few) steps away is it from the existing design? I'm no computer scientist but it feels like the Greeks would surely have figured it out pretty quick (after one of them put in a PRD for a general arithmetic calculator of course).
It's an analog computer.
One important tangent from this and the other ancient wonders is the peril of proprietary information.
How much earlier might humanity have achieved calculus, flight, transistors, &c if important knowledge wasn't confined to priesthoods and artisans?
Don't let the know-how pass with you, say I.
I don’t think the ancients were altogether different from us when it comes to protecting trade secrets. Innovations of military or economic value were kept secret to preserve an advantage, and they still are today.
I also don’t think there’s a slam dunk answer to whether innovation is spurred more by the profit motive inherent in the proprietary approach, or by the widespread enablement of the high-openness approach.
Yes. Secrecy associated with silk manufacture is a great example [0].
[0].
Unsure if you missed pasting the link, or you're being ironic...
This is amazing: denote empty citations as an alernate to the terminating sarcasm tag (/s). It's probably even more explicit, and we are forced to charity by the absence of reference.
Imagine what wonders the wandering of undeclared references shall do to a mind numbly scrolling through comments; forced to pause, and brief attention focused on the fitting of a slotted thought.
The historical record seems to show a great deal of knowledge being obtained, then lost.
Knowledge transfer ain't bean-bag.
Calculus, flight, transistors etc. were not delayed because of knowledge being confined, but because of knowledge being persecuted.
(a) Both can be among the causes for slow knowledge growth
(b) The past tense is by no means in order where persecution is concerned.
Sure, agreed with (a) in general that they both can be causes of decline, but in the specific case of this post, it is amply clear that the wide range of intellectual areas that ancient Greece excelled in (with barely a parallel) was radically narrowed and almost destroyed due to persecution of this knowledge rather than it being restricted to the elite.
I don't understand what you're trying to say with (b).
(a) Euclid and Pythagoras have surviving writings, but the physical applications were lost.
(b) Look at the scientists trying to view Covid accurately.
What I'm missing, and we all should, are all the other mechanisms - whatever you call them computers or not - which came before this one. Or after. Because it certainly didn't just pop into existence but was the product of a long school of mechanism builders. Where are they? And why no more complicated ones were found? Or later ones? Wasn't it enough successful? Why?
I imagine it's a similar story to why the industrial revolution did not occur in ancient times, despite elaborate steam powered machinery being known technology: insufficient demand to promote scaling the technology into affordability / practicality. Roughly speaking I imagine it took a master craftsman on the order of a year to make the Antikythera mechanism, and all it did was automate a fairly niche task that could be done relatively easily by hand anyway. It is fascinating to consider that there must have been many similar devices that we've never found, and tantalizing to wonder if there are other classes of ancient mechanical computer we're not aware of. But fundamentally, they're very very expensive toys, not ubiquitous household items like coins. Even if a thousand were made, which would take generations to do by hand, we are still lucky to have found even just one.
Nearly 2000 Albatros D.III fighter aircraft were built to fight in WW1. Today, only tiny fragments survive in museums around the world. What will be left in two thousand years?
I got to see this the other day!
In the archeological museum in Athens, they have the recovered pieces, and a number of videos and modern replicas to try and explain how it works. It was very cool to see something like that in real life - not only the mechanism representing a tremendous amount of skill and craftsmanship, but the knowledge of how to precisely measure out decades-long patterns of celestial movement and fit them all together. Things like figuring out a slight irregularity in the pattern of the moon, and then figuring out a way to put gears together to replicate the same irregularity - and creating a device that would have been usable centuries after their lifetimes (if it hadn't sunk to the bottom of the sea, of course!)
That said, the most exciting thing for me to see there was definitely the Linear B tablets, with some of the earliest texts on them that we can understand. All in all, a wonderful collection of artefacts.
It's an amazing piece of technology, but I read somewhere that the engineering tolerances of the gears would have caused drift/errors with the outputs that would have made it unusable without being recalibrated.
I don't know if this would really address your comment, but clickspring [1] on YouTube has a video series creating a modern replica of the mechanism with period appropriate tools.
[1] https://youtube.com/playlist?list=PLZioPDnFPNsHnyxfygxA0to4R...