by Decoding the Heavens- Solving the Mystery of the World's First Computer (retail) (epub)
He knew that his best chance – of cracking the mechanism and making a compelling film – lay in getting the best possible images of the gearing. Price’s radiographs were lost. Freeth had heard from Mike Edmunds that some chap from the Science Museum had also X-rayed the fragments, but that was a decade ago, and as far as Freeth could tell nothing had ever come of it. Probably nothing ever would, he thought dismissively. From what he knew of the crude tomography technique the man had used, his images would be fiendishly difficult to interpret. Freeth wanted something altogether more sophisticated.
He scoured the scientific literature for reports of the latest imaging technology, and found two articles that promised a way forward. First, he saw on Nature’s front cover a colourful picture of a goldfish’s innards, showing its feathery ribs and even its internal organs in subtle yet sharp detail. Normally X-rays are blocked by bone but pass straight through a body’s soft tissue, which is why a conventional X-ray image of a person shows just their skeleton. This paper’s Australian authors had perfected a technique that detects the change in the phase of X-rays (how the highs and lows of the different waves line up) as they pass through different types of material, achieving intricate images of objects that wouldn’t normally show up at all.
Freeth emailed the researchers to ask whether the method would work on the Antikythera mechanism. But the reply came back explaining that unless they were chopped into thin slices, the dense bronze of the fragments would likely block the X-rays altogether. Instead the authors suggested that Freeth investigate a technique called microfocus X-ray imaging, which uses a powerful but tiny X-ray source so that the radiation it emits can be focused very precisely to give a high-resolution image. The latest machines used digital detectors rather than film, which recorded the precise amount of radiation hitting each pixel and fed it directly into a computer.
Microfocus X-ray imaging sounded like the perfect way to gain clear images of the tiny inscriptions and gear teeth of the Antikythera fragments. So Freeth found his way to Hadland’s X-Tek. The company’s X-ray sources were only a few thousandths of a millimetre across, yet capable of producing strong beams of radiation. Freeth spoke to the sales manager, who agreed to help almost immediately. The project sounded straightforward – well within the capabilities of X-Tek’s equipment – and the press coverage the results were likely to get would be a great marketing opportunity for the company.
The second article, in New Scientist magazine, described a technique that was being used to decipher some of the oldest writing in the world. These were cuneiform inscriptions; the wedge-shaped letters that the people of Mesopotamia – today’s Iran, Syria and Iraq – scratched into wet clay as far back as the fourth millennium BC. The surviving tablets, though amazingly well-preserved for their age thanks to the dry sand that has cradled them for so long, are worn and faded, with many of the inscriptions now quite illegible. The article described how a young researcher working at Hewlett-Packard’s labs in Palo Alto could transform photographs of these dull tablets into sharp, glossy computer images with inscriptions that practically jumped off the screen.
The researcher was a floppy-haired Californian called Tom Malzbender, and decoding invisible messages from past civilisations was the last thing that he had expected to find himself doing. He had actually been trying to develop more realistic computer graphics. It was easy enough to create a computer image of an object, a knight in shining armour, say, and have it clank around the screen. You just wrote a piece of code to model the texture of the material you wanted, then wrapped it round the appropriate geometric shape for each frame of the animation. But for a more realistic effect, Malzbender wanted to model how different materials look as they move past a light source. The reflections on that suit of armour will change as the knight walks under a chandelier or steps out into the sunlight. And the changing appearances of materials that reflect light in more complicated ways – curly hair, for example, or a crumpled newspaper – are even harder to calculate.
Tom Malzbender’s relaxed exterior hides an acute technical expertise, combined with the sort of mind that refuses to see the world according to normal rules. Instead of trying to come up with more sophisticated mathematical models, as his competitors were doing, he decided to let nature do some of the work for him. He realised that if he took lots of digital photographs of the same object lit from slightly different directions, he could feed those images into a computer and measure how each pixel in the picture varied – creating a virtual map of the object’s response to light. So he built a light-proof dome, with a camera fixed inside at the top, pointing downwards. The rest of the dome’s inner surface was covered with 50 flashbulbs, wired to fire one after the other, each time the camera shutter flicked open. The equipment looked like a home-made flying saucer, but it worked just as he hoped. Once the computer had done its analysis it was like having a magic wand that imposed his will on to the screen – he could dim the lights, turn them up or wave a spotlight over the object in the image with the twitch of a mouse. He could even create effects that would be impossible in the real world, such as lighting a surface from behind, or hanging a lamp inside the thinnest scratch.
In 1999 Malzbender attended a lecture given by an archaeologist who was trying to decipher ancient cuneiform tablets. Realising that his imaging technique might help to enhance the faded inscriptions, Malzbender offered his services and was given a crumbling tablet on which to test out his idea.
After photographing it, he played around with the resulting image. As well as changing the position and brightness of the light source, he realised he could ask the computer to change the way that the object itself reflected light. For example, making every pixel reflect light more strongly worked like coating the dull, dusty surface in glossy metal, making every defect as obvious as a scratch on a brand-new car. And by setting each pixel to only reflect when the light was pointing directly at it, he could make tiny marks jump out like gleaming stars against the dead black of night.
The results were stunning. The tablet was a draft contract, written around 3100 BC for a Sumerian slave trader called Ur Ningal. In Malzbender’s images the inscriptions on the worn, crumbly surface appeared as clear as shining crystal, including some that it hadn’t been possible to read at all before. He even saw the fingerprints of the scribe who held the clay while it was still wet.
As soon as Tony Freeth saw the pictures he wanted the dome for his Antikythera project. He e-mailed Malzbender in 2001 to explain his idea, but the graphics expert didn’t think much about it – he received a lot of requests to use his dome to image various objects, and most of them never came to fruition. The next year, however, Malzbender happened to spend three months on sabbatical in Bristol and while he was in Britain he travelled to London to meet Freeth for lunch in the refined surroundings of the National Gallery.
Malzbender was keen to work out whether Freeth was serious about the proposed project. The man seemed quiet, almost awkward – very British, Malzbender thought. But five minutes into their conversation, he was impressed: he was used to explaining the technicalities of his algorithms over and over to duller minds than his own, but Freeth understood every detail instantly. Ten minutes in, there was nothing more to say. Malzbender knew that he would be going to Athens.
Freeth now had two companies willing to help him get more information out of the Antikythera fragments than ever before. X-Tek’s state-of-the-art technology promised to make sharp X-ray images of the gearwheels inside the device. And Hewlett-Packard’s revolutionary light mapping would help him to read previously hidden surface inscriptions.
But there were two big problems. Freeth didn’t have any funding for the project; although both companies had promised to give their time for free, he still needed to raise enough money to ship all of the necessary people and equipment to Athens. Worse, he didn’t have permission to study the fragments.
Officials at the Athens National Archaeological Museum had turned down his request. A campa
ign was hotting up to persuade Britain to return the Elgin marbles to Athens in time for the 2004 Olympics there, so a British researcher wanting access to a Greek artefact was not in a good bargaining position. But Freeth would not take no for an answer. He calculated that if he could get some prominent Greek scientists on board the museum might change its mind, so he set about putting together a collaboration that would be influential enough to push his plans through.
Trying to solve the mystery of the Antikythera mechanism gave Tony Freeth a sense of purpose that neither his maths nor his films had ever given him, and over the next couple of years all of his other projects fell away as he devoted more and more time to his quest. He organised petitions, published papers and wrote grant applications. He set up an e-mail discussion group devoted to gaining access to the fragments, and lobbied everyone he could think of for support. And he never stopped going over Price’s old work.
Gradually a team came together. The first recruit was John Seiradakis at the University of Thessaloniki, one of Greece’s most eminent astronomers; then came Seiradakis’s friend Xenophon Moussas, an astrophysicist at the University of Athens with close links with the National Museum; and there was Agamemnon Tselikas, director of the Centre for History and Palaeography in Athens and an expert in reading ancient texts. Finally, Mike Edmunds provided the necessary academic credentials from Britain. They were a band of brothers, Freeth thought proudly. He would lead them to victory.
John Seiradakis and Mike Edmunds, as the most senior scientists on the team, applied to the National Museum with the full force of their joint academic reputations. Edmunds also applied for grant money from the Leverhulme Trust, an organisation set up by William Lever (founder of the company that eventually became Unilever) to fund unique and interdisciplinary projects that might not have a chance of support elsewhere. After several unsuccessful, frustrating attempts the team finally won its money early in 2005.
Just two weeks later, the answer came back from officials at the National Museum. Despite the presence of the Greek collaborators, the answer was still no. As far as the museum was concerned, the Antikythera fragments had already been X-rayed – twice – and work on the latest round of data was still in progress. There was no reason to put the crumbling pieces at further risk by imaging them again.
Freeth refused to consider defeat, so he changed his plan of attack. The only organisation with the power to override the National Museum’s decision was the Greek Ministry of Culture. Xenophon Moussas took over the fight.
Moussas is a gentle, soft-spoken man, not normally the type to make trouble. But he is fiercely proud of his Greek heritage. When he was a child growing up in Athens he used to love going to the National Museum – just as a young Michael Wright had been enchanted by London’s Science Museum. Moussas never tired of telling Freeth and the others how he used to stand in front of the Antikythera fragments in their glass case, marvelling at the sophistication of the ancient mechanism and losing himself in daydreams about his forefathers, the ancient Greeks. Away from the museum, he’d look up at the night sky and imagine those astronomers from long ago who were equally inspired by the same sight.
As Moussas grew up, he never lost his fascination with the skies. He became a physicist, specialising in studies of the Sun. Many of the projects he worked on recalled the Greeks’ past glories, from NASA’s Ulysses spacecraft on its distant wanderings over the Sun’s poles, to Artemis 4, a radio telescope positioned in the mountains at Thermopylae, where King Leonidas and his small band of Spartans had once fought the full force of the Persian army.
Moussas took his new role on the Antikythera team seriously. If only he could speak to the Culture Minister he could explain the merits of the project – the quality of the equipment and the eminence of the researchers who were on board, and the importance of understanding the mechanism for Greece! Only the Antikythera mechanism could show the true extent of the ancient Greeks’ achievements – in science and technology, not just art and battle. If the minister would only hear the story, he would surely force the museum to give them access.
Moussas phoned the Ministry of Culture again and again, but he could never get anyone to put him through. Perhaps his persistence eventually wore the secretaries down, however, because after 40, 50, maybe even 60 calls, he was finally granted a meeting with the Deputy Culture Minister Petros Tatoulis and his wife Sofia.
Moussas’s words fell on fertile ground. As well as a passion for archaeology (since his appointment in 2004 Tatoulis had been leading efforts to pressure the British Government into returning the Elgin marbles), it turned out that the couple had a keen amateur interest in ancient astronomy. Not to the entire delight of the National Museum staff, Tatoulis arranged a two-week slot for the team to study the fragments in September 2005. Everything was in place. With just a few months to wait, an elated Tony Freeth told his collaborators at X-Tek and Hewlett Packard to prepare for the trip to Athens.
It wasn’t to be quite that simple. X-Tek’s technology had moved on since Freeth originally approached the company in 2001. Roger Hadland’s engineers had developed equipment capable of combining microfocus X-ray imaging with a technique called computed tomography, or CT. This is basically a much more sophisticated version of the method Michael Wright had used with his home-made tomography cradle. Rather than taking a series of flat radiographs, CT can produce a three-dimensional reconstruction of the entire object, so with the right computer software you can fly right through its insides, exploring every hidden corner. The technique uses an X-ray beam that spreads out from the tiny source in a cone shape, before passing through the target object and then hitting a square detector. Each pixel of the detector measures the precise amount of radiation that hits it. From this, you can draw a series of straight lines through the object, from the X-ray source to each individual pixel, and work out exactly how much radiation was absorbed by the object along each line. That doesn’t tell you much by itself. But then you rotate the object a tiny bit and do the same thing again, in fact you do it thousands of times until you have an image from every possible angle.
Depending on the position of different structures inside the object, each image taken as the object is turned will cause a slightly different pattern of radiation to hit the detector. And that gives the computer enough information to work out the exact arrangement of the object’s insides – like filling in the squares in a game of X-ray sudoku.
Freeth had realised that three-dimensional CT would be the ultimate tool for understanding the internal details of the mechanism, and when he asked the sales team if X-Tek could use it to image the Antikythera fragments, they had said yes. But with the project now actually about to happen, Roger Hadland himself finally got to hear of the promises his team had made.
He was immediately concerned. The Antikythera fragments were made of bronze, a very dense metal, and the fragments were large by the standards of microfocus CT – the largest was more than 18 centimetres across. With normal X-ray images it doesn’t matter if some parts of the target object block X-rays completely, they just come out black in your final image. But with CT, the team would need information from every pixel, even when probing the fragments end on, or the computer wouldn’t be able to accurately reconstruct the internal details.
X-Tek’s microfocus sources work on the same principle as larger ones – an electron beam hitting a tungsten plate knocks electrons in the tungsten atoms up to higher energy levels, which then relax and spit out high-energy photons – X-rays – in the process. The power of an X-ray source is given in volts, which is basically a measure of the strength of the electron beam producing the X-rays. X-Tek’s smallest sources weighed in at 225 thousand volts (225 kV).
No one in the world could better that, but Hadland estimated that to guarantee sharp CT images of the Antikythera fragments would take around 450 kV. Such a machine simply didn’t exist.
Roger Hadland didn’t want his company associated with anything less than perfect data. ‘I’l
l ring Tony and tell him we can’t do it,’ he told the sales team. He had never spoken to Freeth, but he expected it to be a brief call – Freeth would express polite disappointment, Hadland would express regret, and X-Tek’s foray into ancient history would be over.
‘You have to do it!’
It was not a request from Freeth, but a forceful statement of fact. There followed a torrent of admonition, explanation, persuasion and drama the like of which Hadland had never experienced. One hour later, he had made a 180-degree turn. He was now convinced that imaging the Antikythera mechanism would be a once-in-a-lifetime opportunity to discover the secrets of one of the most important artefacts that survives from the ancient world. He was also starting to think about what a tremendous opportunity the project might be for X-Tek.
Imaging the mechanism would require a completely new machine, with X-ray beams twice as powerful as any the company had developed so far. Hadland had long wanted to design such a very high-voltage X-ray source, because it would allow the company to expand into a badly needed new area of business – imaging the turbine blades in aircraft engines.
In a turbine engine, air is heated to searing temperatures by burning fuel, then as it expands it rushes past the turbine blades, turning them and driving the engine. The gases are hotter than the melting temperature of the blades, so they have cooling channels inside to dissipate the heat. If there is any defect in these channels, the blades can fail, with devastating results for the aircraft. If X-Tek could build a microfocus X-ray machine capable of 450 kV, it would be powerful enough to image the tiniest cracks in objects as big as turbine blades.
The problem was, such an ambitious design project typically takes two or three years. Since the bottom dropped out of the microelectronics market, Hadland just couldn’t afford to invest effort in a project for that long without any immediate financial returns. But the more he thought about the Antikythera project, the more his heart ruled his head, and the excitement of invention filled him, as it had when he first built up his company. Freeth had been given a specific time slot at the National Archaeological Museum: September 2005. It was already June, giving Hadland less than four months. It was a ridiculously short space of time, but what if he blitzed it? He could put everything the company could spare into the project for a few months, image the fragments, then dive into new markets with his world-beating X-ray system before the year was out.
