Antikythera Mechanism(en.wikipedia.org)
en.wikipedia.org
Antikythera Mechanism
https://en.wikipedia.org/wiki/Antikythera_mechanism
9 comments
The Clickspring creator hasn't provided updates for months on his Antikythera reconstruction, but recently he updated his patreon supporters of the reason why: He has discovered something new about the mechanism, previously unknown, that he is working on publishing. Fascinating!
Ohh exciting. I saw this thread and wondered what happened to that project.
Cool!
Not only is he recreating the mechanism, he's also exploring antique methods and materials that were possibly used.
On my recent vacations to Athens I visited National Archeological Museum where the Antikythera is displayed.
I highly recomend seeing this device and its recreations live there.
Especially seeing the recontructioned devices gave me completely new perspective about ancient Greece.
Especially seeing the recontructioned devices gave me completely new perspective about ancient Greece.
I had a friend in honours year computer science who did a project on tomography of the antikythera mechanism (working on image analysis software to remove ghost images of other layers), and helped discover why his advisor's previous model of the mechanism was mechanically immobile (when you turned it, it locked). He'd discovered that an axis seen end-on was actually 2 axes, if I remember correctly.
Amazingly enough he forgot to mention this in the end of year presentation of his work, so I got to ask that question to remind him... talk about burying the lede.
Amazingly enough he forgot to mention this in the end of year presentation of his work, so I got to ask that question to remind him... talk about burying the lede.
This guy on youtube is crowdsourcing a series where he is recreating the mechanism https://www.youtube.com/watch?v=ML4tw_UzqZE&list=PLZioPDnFPN...
I like the fact that these devices were well attested in history (https://en.wikipedia.org/wiki/Antikythera_mechanism#Similar_...) but still such a surprise when they appeared in archeology.
It is as if knowledge about them was methodically wiped off history, or more likely (since history is written by non-engineers) it was not understood well enough to write about it.
No need to posit a conspiracy here. People died, and there weren't enough new people entering the field to learn it all before the critical mass of knowledge died with them.
I believe that the bulk of knowledge, in most fields, is passed from person to person through active practice, not through books. Books, as a cheap way to widely distribute knowledge, are relatively recent innovation, and most of what I know about any topic is not written in books. Or if it is, it's almost an afterthought, not the primary method of knowledge transfer. What book did you read to learn to tie your shoes, or drive a standard transmission?
On a related note: what is humanity's Bus Factor for creating a CPU? I know a little about semiconductors from school but I certainly wouldn't know where to start, even for a tiny 8-bit one. I doubt Intel writes down everything they've figured out. Could 2 or 3 especially unlucky "city-killer" asteroids send us back to a pre-information-age society?
I believe that the bulk of knowledge, in most fields, is passed from person to person through active practice, not through books. Books, as a cheap way to widely distribute knowledge, are relatively recent innovation, and most of what I know about any topic is not written in books. Or if it is, it's almost an afterthought, not the primary method of knowledge transfer. What book did you read to learn to tie your shoes, or drive a standard transmission?
On a related note: what is humanity's Bus Factor for creating a CPU? I know a little about semiconductors from school but I certainly wouldn't know where to start, even for a tiny 8-bit one. I doubt Intel writes down everything they've figured out. Could 2 or 3 especially unlucky "city-killer" asteroids send us back to a pre-information-age society?
This seems to happen a lot. One people seem to keep chasing is Damascus steel.
Even in more recent history, too. I like to research fine weapons. Apparently the process Colt used on their pistols to color and shine them is completely undocumented and now unknown. Similarly, it's said there are not enough people alive that know how to handfit a revolver. I'm not sure how true the latter is, but really makes you wish Wiki or the internet existed further back.
Even in more recent history, too. I like to research fine weapons. Apparently the process Colt used on their pistols to color and shine them is completely undocumented and now unknown. Similarly, it's said there are not enough people alive that know how to handfit a revolver. I'm not sure how true the latter is, but really makes you wish Wiki or the internet existed further back.
Also, supposedly, the styrofoam used inside thermonuclear weapons.
What’s your goal? Rebuilding an Intel class modern CPU or just making a crude one?
Transistors give you logic gates and opamps, from which you can build computation and memory, add a clock to apply input from a certain address (the PC) to a set of logic gates (the instruction set) that can read and store and you have a very crude (and very inefficient) programmable CPU in no time (compared to something even remotely like a 8080 or even a 6502)
A good approach to get more knowledgeable is going bottom-up from transistors and top-down from FPGA (and suffer Verilog/VHDL), attempting to experiment and design your own instruction set while trying to understand the physics and logic, progressively bridging the gap in between in both directions.
Transistors give you logic gates and opamps, from which you can build computation and memory, add a clock to apply input from a certain address (the PC) to a set of logic gates (the instruction set) that can read and store and you have a very crude (and very inefficient) programmable CPU in no time (compared to something even remotely like a 8080 or even a 6502)
A good approach to get more knowledgeable is going bottom-up from transistors and top-down from FPGA (and suffer Verilog/VHDL), attempting to experiment and design your own instruction set while trying to understand the physics and logic, progressively bridging the gap in between in both directions.
Relevant book: A Canticle For Leibowitz.
If you’re talking about losing the ability to recreate any particular relevant CPU, then maybe. But computers in general? No. I suppose we could be exceptionally unlikely and lose access to all the modern fabs, but the knowledge of CPU design is too widespread to for us to plausibly lose the ability to make new CPUs altogether.
A whole lot of people know how to make a CPU. It's not as hard as you probably think it is. What's hard is making them as complex, fast, and small as we do nowadays. But if you can get by with something not as complex or fast, and that takes up most of your desk, thousands of people know how to do it.
Got 120 hours and want to become one of them? Take these two courses from EdX, which are 60 hours each:
"Computation Structures - Part 1: Digital Circuits" [1]
"Computation Structures 2: Computer Architecture" [2]
The first teaches "[...] digital encoding of information, principles of digital signaling; combinational and sequential logic, implementation in CMOS, useful canonical forms, synthesis; latency, throughput and pipelining". In the homework and labs you design and implement (in a simulator) a 32-bit ALU.
The second covers "[...] instruction set architectures and assembly language, stacks and procedures, 32-bit computer architecture, the memory hierarchy, and caches". In the homework and labs you design and implement (in a simulator) at the gate level a 32-bit RISC CPU, except for memory. Memory for registers and program is given as a black box--by this point you know enough to design that, but it would just add a lot of components and complexity, and slow down the simulation, and the time spent dealing with it would distract from learning the topic of this part of the course. That would fit better with the first part of the course.
I've taken these, and can say they do a good job of teaching what they say they teach.
There's also a third course in the series:
"Computation Structures 3: Computer Organization" [3]
That covers "[...] pipelined computers, virtual memories, implementation of a simple time-sharing operating system, interrupts and real-time, and techniques for parallel processing".
In the homework and labs for that one, you optimize your CPU from the second part for size and speed, and make it support time sharing operating systems.
I've not taken this one.
[1] https://www.edx.org/course/computation-structures-part-1-dig...
[2] https://www.edx.org/course/computation-structures-2-computer...
[3] https://www.edx.org/course/computation-structures-3-computer...
Got 120 hours and want to become one of them? Take these two courses from EdX, which are 60 hours each:
"Computation Structures - Part 1: Digital Circuits" [1]
"Computation Structures 2: Computer Architecture" [2]
The first teaches "[...] digital encoding of information, principles of digital signaling; combinational and sequential logic, implementation in CMOS, useful canonical forms, synthesis; latency, throughput and pipelining". In the homework and labs you design and implement (in a simulator) a 32-bit ALU.
The second covers "[...] instruction set architectures and assembly language, stacks and procedures, 32-bit computer architecture, the memory hierarchy, and caches". In the homework and labs you design and implement (in a simulator) at the gate level a 32-bit RISC CPU, except for memory. Memory for registers and program is given as a black box--by this point you know enough to design that, but it would just add a lot of components and complexity, and slow down the simulation, and the time spent dealing with it would distract from learning the topic of this part of the course. That would fit better with the first part of the course.
I've taken these, and can say they do a good job of teaching what they say they teach.
There's also a third course in the series:
"Computation Structures 3: Computer Organization" [3]
That covers "[...] pipelined computers, virtual memories, implementation of a simple time-sharing operating system, interrupts and real-time, and techniques for parallel processing".
In the homework and labs for that one, you optimize your CPU from the second part for size and speed, and make it support time sharing operating systems.
I've not taken this one.
[1] https://www.edx.org/course/computation-structures-part-1-dig...
[2] https://www.edx.org/course/computation-structures-2-computer...
[3] https://www.edx.org/course/computation-structures-3-computer...
It’s not the logic but the technology that is not widely known: the new fabs cost billions.
It seems like only a handful of people ever made them and it probably didn’t even last for very long after Archimedes died. It was essentially an extraordinarily expensive toy.
Because the design and engineering greatly outstripped the extant methods of manufacturing the devices didn’t actually work that well.
It would take another 1000 years of slow incremental improvements in manufacturing to be able to make one of these devices accurate enough to be useful.
If you want to make civilization resistant to collapse and increase its ability to recover quickly: bury gear cutters, lathes, magnifying glasses, precision tools, etc. The most difficult part will be bootstrapping the ability to build accurate tools with which to rebuild an industrial society.
It would take another 1000 years of slow incremental improvements in manufacturing to be able to make one of these devices accurate enough to be useful.
If you want to make civilization resistant to collapse and increase its ability to recover quickly: bury gear cutters, lathes, magnifying glasses, precision tools, etc. The most difficult part will be bootstrapping the ability to build accurate tools with which to rebuild an industrial society.
One of the things most interesting to me about the mechanism is that it devotes a lot of gears and hardware to making small corrections which get completely overwhelmed by the inaccuracy of the manufacturing.
It's actually possible to build a working machine lathe out of the parts of an ordinary car engine. If you really wanted to make society collapse resistant, including a manual about how to do so, a couple of tungsten carbide bits, and a couple of hacksaw blades with every car would probably do the trick. I think everything else you need (welding equipment, drills, etc) would be fairly straightforward to cobble together.
It was essentially an extraordinarily expensive toy.
So not unlike early cars.
By 1900, at least 100 different brands of horseless carriages were being marketed in the United States. Since they were all virtually handmade, the cars were outrageously expensive. Cars were perceived as no more than a high-priced toy for the rich. The early 1900s cars were, to many, a despicable symbol of arrogance and power.
https://www.anythingaboutcars.com/early-1900s-cars.html
So not unlike early cars.
By 1900, at least 100 different brands of horseless carriages were being marketed in the United States. Since they were all virtually handmade, the cars were outrageously expensive. Cars were perceived as no more than a high-priced toy for the rich. The early 1900s cars were, to many, a despicable symbol of arrogance and power.
https://www.anythingaboutcars.com/early-1900s-cars.html
There's good evidence that this kind of thing was made in the Arab world in the "dark ages" and that many of the moving statues in western europe were powered from the same sort of mechanisms in the medeival period. Great listen here : https://www.bbc.co.uk/programmes/b0bk1c4d
It could be that the thing just had no use, or that it became obsolete. I can imagine prediction of the seasons being useful for agriculture and trade, but people may have figured out how to do that without a calculating machine. Astrological uses would have depended on the particular astrology being in use.
Information from antiquity is lost by default. It is only preserved if there has been is a continuous line of people preserving it.
I think - just not of interest !
Plenty of previous discussion: https://hn.algolia.com/?query=Antikythera%20comments%3E0&sor...
And how does posting a Wikipedia article link make it to near the top of Hacker news?
Firstly someone has to create a mechanical calculating technology that is lost to history before being discovered in a shipwreck by a society that has globally networked its modern equivalent. After that it gets complicated.
As long as they're interesting Wikipedia submissions, they are on topic. Why wouldn't they be?
For people that want to try this with a good quality 3d printer:
https://www.thingiverse.com/thing:1821540
I recommend 0.1 layer height and conservative settings for a slow, but high quality print.
https://www.thingiverse.com/thing:1821540
I recommend 0.1 layer height and conservative settings for a slow, but high quality print.
Jonathan Blow draws some interesting implications of discoveries like this one in a recent talk: https://www.youtube.com/watch?v=pW-SOdj4Kkk
Any 3D printed versions?
There is a really cool LEGO version someone built and you can see on YouTube.
https://www.youtube.com/channel/UCworsKCR-Sx6R6-BnIjS2MA