Mathematical and scientific inventions have scored the history of mankind just as the great poems of the West and the East have left their indelible mark. These marvels tell our shared story, but they would be nothing without the paper they’re written on. That may seem trite but it’s not. Let me tell you why, with three examples that show how important it is to store information.

The first is the Iliad and the Odyssey. Writing was not widespread in Greece until the sixth century BC, when, fortunately, some unknown bright spark had the idea of setting down in black and white the poems attributed by custom to Homer. Whoever that was deserves a posthumous Nobel Prize.

The second example is less fortunate and concerns the work of one of the fathers of Western thought, Heraclitus. Fragments are all that remains of his On Nature, penned between the sixth and fifth centuries BC. We know that a copy of the book was kept at the library in Constantinople, but it went up in flames during one of the Crusaders’ futile slaughters in the Byzantine capital.

The third example was lucky, however – very lucky. Archimedes is well known for coming out with some good ideas. The Arabs of the early Middle Ages credited him with an integral calculation method almost two thousand years before the first coherent definition of an integral of a mathematical function, in the second half of the 17^th century. Archimedes wrote down his calculation system, along with others, in a weighty scroll that ended up in Constantinople, then Jerusalem, where it was washed and reused by a copyist for a prayer book. That might have been the end of Archimedes’ integral calculation, had it not been for a Greek scholar in the late 19^th century who made out the earlier text written backwards underneath the prayers. Just a century later, in 2008, Stanford University’s laboratories finished the painstaking work of drawing out everything the old mathematical genius had written.

For many centuries, until at least the beginning of the 19^th, what we now call memory – that is, the memory outside of us, not within us – was preserved on stone, papyrus and paper, with the chisel, the awl and the quill, the ancestors of the fountain pen and László Bíró’s ingenious invention. Tables, rolls, books, textbooks and newspapers. That’s how memory was stored. Information was kept in the same language we used to speak. This was memory written and made of words, sentences, speeches, concepts, theories, true stories and tall tales, information.

But when Joseph Marie Jacquard, a weaver from Lyon, decided to create a series of instructions for looms, to make them do more complicated things, the book, textbook and newspaper were of no use to him. To automate the instructions, he had to transform and codify them, make them capable of giving orders to a machine. When in 1801 Jacquard revealed his contraption, the first textile loom to produce complicated designs without the help of an operator, it was accompanied by an erudite description written in pen on paper. To present his invention, he relied on two memories, one mental, transcribed on paper, the other made of a roll of holed card. The former explained his machine, but the latter made it run, letting needles enter the holes, “read” the instructions and pass them on to the heddles making the stitches in the cloth. Jacquard’s sheets had a double function: they remembered the design to be stitched and the instructions for the machine.

In a way, that links them with our computers. Of course, the programs that let us write, do calculations, draw and do so much else on our computers are distinct from the things we produce with them, but both are conserved in the machine’s memory. Now, the more advanced the program and the things it produces, the more memory is needed to store them. So, when the United States and Britain invented the first computers, mainly to crack coded information sent by German military commanders, they looked to Jacquard’s idea, the holed sheets that could store relatively large amounts of information and read it without any ambiguity. One sheet made by IBM in the 1950s, a rectangle of cardboard with 80 columns and 12 series of holes, could store information equivalent to 128 bytes.

Not many years later, UNIVAC was launched, the first computer with memory on magnetic tape, with a storage capacity of no less than 225 kilobytes, the equivalent of almost 2,000 holed sheets. But the tape had a weakness: the information on it could only be read in sequence. The solution was the hard disk, from which data could be summoned in any order. The first was made by IBM and had a capacity of 5 megabytes. Next in line was the floppy disk, for transferring information from one machine to another, with a capacity of 1.44 megabytes, followed by CD-ROMs and DVDs, with capacities of up to 5 gigabytes. The tide of 40 years eventually brought the modern USB, which can hold up to 128 gigabytes, the exact equivalent of 1 billion holed sheets.

It’s an extraordinary result. And it’s not the end. Computing power has gone up so much that it needs even more copious memories, which must also read and write data faster. Building them isn’t an insurmountable problem. Today’s supercomputers can have a memory equivalent to many, many billions of billions floppy disks and are managed by reliable, lightning-fast software. So, all good? Not quite. One problem remains: the fragility of information. It can be destroyed in a fire, in a flood, any event strong enough to damage a computer and its memory system beyond repair.

The solution so far has been to back up the information on other computers. But storing this data over time is still tricky. UNESCO has come up with a solution, the “Software Heritage” project, to protect data and the programs it’s written on, so no memory from the past is lost. That goes for books, the things we guard so jealously in our libraries. Common sense dictates we should do the same with the information on our computers. We can’t rely on the kind of luck Archimedes’ calculation process had to save the past from oblivion.