Antikythera: 2000 year old computer reconstructed

[Xeraux]

Originally Posted by Klay

I've read some other stories about him. I was going to trot him out at some point but now you've pre-empted me.

Mad semantic skillz, Baby.

[Nixels]

Originally Posted by Xeraux

The Professor and the Madman: A Tale of Murder, Insanity, and the Making of The Oxford English Dictionary was about Dr. W. C. Minor who contributed thousands of definitions from his cell in an insane asylum.

It was a good read.

That is a great read.

Thanks for taking up the semantics battle while I was drinking coffee, Xeraux.

Now I'll have to take back all that shit I slung around about you when I was hanging out with Hedgie in the frozen northlands.

[Albert]

Originally Posted by Klay

Well sure, a word means whatever we mean it to mean.

Don't they really need punch cards and vacuum tubes to really be computers?

[Klay]

Originally Posted by Albert

Don't they really need punch cards and vacuum tubes to really be computers?

No, not at all. They can be completely mechanical and made out of materials such as wood.

[olebiker]

Originally Posted by Nixels

That is a great read.

Thanks for taking up the semantics battle while I was drinking coffee, Xeraux.

Now I'll have to take back all that shit I slung around about you when I was hanging out with Hedgie in the frozen northlands.

Any idea who the author is?

[Klay]

Okay, I've got it. It's an Orrery. From Wikipedia:

According to Cicero, the Greek philosopher Posidonius constructed an orrery, possibly similar or identical to the Antikythera mechanism, which exhibited the diurnal motions of the sun, moon, and the five known planets. Cicero's account was written in the first century BC.
The Antikythera mechanism may be considered one of the first orreries. It is an ancient mechanical calculator (also described as the first mechanical computer) designed to calculate astronomical positions. It was discovered in an ancient shipwreck off the Greek island of Antikythera, between Kythera and Crete, and has been dated to about 150-100 BC. Technological artifacts of similar complexity did not appear until a thousand years later.
The first modern orrery was built circa 1704 by George Graham. Graham gave the first model (or its design) to the celebrated instrument maker John Rowley of London to make a copy for Prince Eugene of Savoy. Rowley was commissioned to make another copy for his patron Charles Boyle, 4th Earl of Orrery, from which the device took its name. This model was presented to Charles' son John, later the 5th Earl.
Joseph Wright's picture "The Orrery" (ca. 1766) which hangs in Derby Museum and Art Gallery, features a group (three men, three children, and a lone woman) listening to a lecture by a 'natural philosopher'—the only light in the otherwise darkened room is apparently from the centre of the brass orrery, which, in the case, has rings that cause it to appear to be similar to an armillary sphere. Shoemaker John Fulton of Fenwick, Ayrshire, built three between 1823 and 1833 - the last is in Glasgow's Kelvingrove Art Gallery and Museum.

[edit] Explanation

Orreries are sometimes called planetariums, although this word usually refers to hemispherical theatres in which images of the night sky are projected onto an overhead surface. Orreries can range widely in size from hand-held to room-sized.
Orreries are usually not built to scale. Some fixed Solar system scale models have been built and are often many kilometres in size. An innovative concept is to have people play the role of the moving planets and other Solar system objects. Such a model, called a Human Orrery, has been laid out with precision at the Armagh Observatory.
A normal mechanical clock could be used to produce an extremely simple orrery with the Sun in the centre, Earth on the minute hand and Jupiter on the hour hand; Earth would make 12 revolutions around the Sun for every 1 revolution of Jupiter. Note however that Jupiter's actual year is 11.86 Earth years long, so this particular example would lose accuracy rapidly. A real orrery would be more accurate and include more planets, and would perhaps make the planets rotate as well.
Many planetariums have a projection orrery, which projects onto the dome of the planetarium a Sun with either dots or small images of the planets. These usually are limited to the planets from Mercury to Saturn, although some include Uranus. The light sources for the planets are projected onto mirrors which are geared to a motor which drives the images on the dome. Typically the Earth will circle the Sun in one minute, while the other planets will complete an orbit in time periods proportional to their actual motion. Thus Venus, which takes 224.7 days to orbit the Sun, will take 0.7 minute to complete an orbit on an orrery, and Jupiter will take 11.86 minutes.
Some planetariums have taken advantage of this to use orreries to simulate planets and their moons. Thus Mercury orbits the Sun in 0.24 of an Earth year, while Phobos and Deimos orbit Mars in a similar 4:1 time ratio. Planetarium operators wishing to show this have placed a red cap on the Sun (to make it resemble Mars) and turned off all the planets but Mercury and Earth. Similar tricks can be used to show Pluto and its three moons.

[Klay]

Originally Posted by olebiker

Any idea who the author is?

http://www.amazon.com/Professor-Madm.../dp/006099486X

[LuciferMutt]

Originally Posted by Klay

No, I'm not thinking digitally. Isn't what makes a computer not just a calculator its ease of reprogramming? This is a definitions question.

No. A computer carries out computations. Some can do many, some can do one. Programming has nothing to do with it. And a calcuator is a computer.

[olebiker]

Originally Posted by Klay

http://www.amazon.com/Professor-Madm.../dp/006099486X

Thanks Klay.

[Putts]

So if I asked the gizmo in question what positions of shit was on such sand such a date, you'd have to spin the no a bazillion times an run it forward to that date to get the answer?

If so, I'd say it was an 'analog' and not a computer.

As someone said before, it ain't reckoning.

[Rafiki]

Originally Posted by Klay

That could be, but then why don't we call a clock a computer?

Because it is a single function computer.

think time computing device.

Originally Posted by Putts

If so, I'd say it was an 'analog' and not a computer.

analog vs digital.............thats just how it does the computing.

analog speedometer vs digital speedometer.

still a speedometer.

[Klay]

It's a gadget that mimics the motions of the heavenly bodies. That's all it does. When it's called a "computer," people transpose their idea of a modern computer...a device with almost unlimited flexibility for managing and exchanging information and performing calculations...on to the ancient orrery, and may come to the conclusion that the ancients had what they didn't. That's misleading.

[Bake]

Originally Posted by Klay

It's a gadget that mimics the motions of the heavenly bodies. That's all it does. When it's called a "computer," people transpose their idea of a modern computer...a device with almost unlimited flexibility for managing and exchanging information and performing calculations...on to the ancient orrery, and may come to the conclusion that the ancients had what they didn't. That's misleading.

A decent CrAsH imitation. You made my list.

[Klay]

Originally Posted by Bake

A decent CrAsH imitation. You made my list.

[Putts]

Originally Posted by Klay

I want to call the Antikythera an astronomical clock. A position calculator for the heavenly bodies. If it can't be easily adapted to calculate the rate grain rises in a bin when fed by a conveyor belt running at varying rates I wouldn't call it a computer.

I don't know that a computer or calculator needs to be able to solve every type of problem. I think I'd be comfortable calling it an astronomical computer if you could set the current date and the desired date; turn the crank a few times; and come up with the answer. Then set the thing back to the current date, set a significantly different desired date; crank the knob about the same number of times; and come up with the answer.

I guess my definition of a computer is when the gizmo executes a calculation that is markedly different and provides a short-cut from merely replicating the physical process in miniature to get answers.

It's still a hell of an achievement, though.

Originally Posted by Klay

I remember the mechanical "target data computers" that were used during WW2 to calculate firing solutions for weapons. It's true I tend to thing of these kind of devices as calculators because the "software" is not easily changeable. I think of a computer as the type of machine which can be told to calculate entirely different types of problems easily..."reprogrammed."

Well, Klay, I know a heck of a lot about that, I was in the Coast Guard in the '70s and was a Gun Fire Control technician. We had a WWII vintage gunfire control system which I worked on.

The computer it self was a roughly ice-chest sized enclosed and seal box in the middle of the gun fire control room.



Inputs came from numerous places: ships bearing, pitch, roll, and yaw came from gyros in the form of synchro/servo signals and continually drove shafts in the computer. We had a couple of refrigerator sized racks of tube electronics amplifiers that acted as the syncro/servo signal amplifiers.

We also got inputs from the 'director', which is the box that had the radar dish on it and would track and lock on to an air/sea/land radar target.

Here's one like we had, but on a WWII Navy ship.



The director had two seats in it and you could drive it with a control stick until you aimed at the target though it's optical system. Once you were on target, you pulled a trigger and the electronic tracking system would lock on to the target and track it. It (and it's associated electronics in the fire control room below decks) sent range, bearing, and elevation to target to the computer.

There were also various temperature and wind information sent to the computer from the nav systems on the boat; and the gun mount itself sent information about the round being used---weight, charge, etc.

The computer had all these little sight holes in it, and during war games when we had a target locked in rough seas, that thing was a liquid blurring whir of gears, cams, and shafts.

Just one of the myriad of calculations that it would make is the steering effect of gyroscopic precession as the projectile when up and over the apex of its trajectory. Because the bullet is spinning (clockwise as viewed from the rear) as it goes up and over the top of its path, it would not only go from being pitched up to being pitched down, but it would yaw to the right, which would steer the projectile to the right. The amount of steering---and therefore compensation needed to hit the target---would be based on factors like gun elevation angle, projectile weight, projectile air resistance profile.

Factors like spin rate due to gun barrel rifling pitch, and offsets developed in bore sighting procedures done in dry dock were dialed in with shaft settings and locked down with a nut.

Here's a pic of a test bed for developing gun fire control computers.






Here's some cool WWII PDFs on the mechanical sub-systems of a basic analog gun fire control computer circa 1944.

Good stuff.

(PS Heres a bairly related but friggen cool page I stumbled across in the search ... have fun.)