Quibble quibble. First, there is no guarantee at all that any particular sequence

*is* in the digits of π, unless you've already calculated that far and found it. A common misconception, but see

**is pi normal?** for the current state of play. So we can't be sure we'll find the

Gödel number of the

*Britannica* in that particular representation.

So let's store it as the data of a quantum computer. Toothpicks are made of wood: what's wood made out of? Carbon, oxygen, hydrogen, a few other elements. And how much does one weigh? Let's use a big chunky toothpick that contains a whopping 2 g of atoms the average size of carbon atoms: that's one sixth of a mole, and therefore contains 1/6 x Avogadro's number of atoms. That'll do for a round number: 10^{23} atoms.

Call the content of the *Britannica* a round thousand million bits. Then each atom has to store 10^{14} bits, or about 2^{50}. A much more manageable number.* All we need now is a bit of quantum superposition

*Read footnote, not this next bit...*
Hmmm... tricky. A carbon atom's got about twelve nucleons and twelve electrons, oxygen a few more, but on average about 25 particles per atom. Each of these might have a few discrete states it can be in, like spin, but I foresee falling short if we rely on those. Let's instead just consider the six to eight shell electrons on each atom... say 8 because I can divide it into 2^{50}.

So we need to excite every electron in the toothpick simultaneously into the first 2^{47} or so energy levels above their ground state.

* Hey, tdent points out an even easier way, which involves *not getting the maths wrong*. 10^{23} atoms storing 10^{9} bits is a whopping 10^{14} atoms per bit. So we can pick and choose which of them get used in the superposition, and maybe even build a detector out of the rest.

As tdent said, the paper *Britannica* itself is not many orders of magnitude larger than a toothpick. Let's say it's 10 000 times bigger in round numbers, 20 kg as against 2 g. This prompts the thought of whether we could actually keep the text on the paper copy: how much heavier is a sheet of printed paper compared to the same blank? A ten-thousandth more? Throw away the heavy covers. Keep just a representative portion (say a central line preserving the shape) of the printed letters, attached to a continuous carbon nanotube: if that's thin enough it should no more than double the mass. Then wind this around the toothpick.

Nanotubes, did I say? You could probably do it in DNA and set up your own miniature printing press.