All this took place quite a few years ago, just after the first confirmed detection of SETI communications. As we all now know, this was kept top secret at the time, although exactly why remains something of an official mystery. I suspect it was the just habitual paranoia of certain government intelligence agencies which made this the default response.

Although I could be wrong. After all the Search for Extra Terrestrial Intelligence had been going on for several decades without a definitive answer and I'm sure the spooks in certain secret-squirrel organisations had been making plans for alien contact for all of that period, if not longer. We'll never know the truth.

Anyway, the SETI signals were confirmed as genuine once a - with hindsight, blindingly obvious - mathematical interpretation had been deduced. Although, the extremely weak signal levels meant that it was hard to be sure at the time. The incoming signals were almost entirely masked by noise, just detectable against the cosmic background - basically, the original radiation of the universe, the relics of the big bang itself.

The appalling signal-to-noise ratio which would inevitably be perceived by any civilisation receiving these signals had evidently been understood by the Originators - the as-yet unknown beings who had transmitted this message. The transmission used an extremely low data rate, so it had taken several years to accumulate enough of the message in an unambiguous form, repeated sufficiently often so that any transmission errors could be cancelled out. This also meant that the signal's artificial nature was confirmed, as it really did repeat precisely every ninety-six-and-a-bit days.

Once the data stream was confirmed, the mathematical properties could be determined - code-breaking, basically - and a plausible interpretation derived. I understand that the exact contents are still a closely-guarded secret, more as a protection against those inevitable pranksters producing a spoof message. Not that it stopped them trying of course.

But the message basically said: "Hi, we're here. We'd like to help. If you'd like to know more, then please build a better receiver." This was followed a large set of what turned out to be technical specifications, the complexity being that we first had to work out the system of units involved. This, it was eventually deduced, was related to certain fundamental units and properties of the universe: the speed of light, the Fine Structure Constant, and so on.

Then followed a theoretical evaluation of the possibility of building what turned out to be a big, really really big receiver, the answer to which turned out to be, perhaps surprisingly, yes, but only just. So the problem turned into one of large-scale engineering and the management difficulties in getting it all delivered. Which is where I got involved, of course.

The general shape of the solution was clear enough: basically, an implementation of very long baseline interferometry, the baseline in question being the diameter of the earth's orbit around the Sun; if you like, an antenna with the resolving power (if not the sensitivity) of a parabolic reflector 186 million miles across. The trick was to get this virtual dish pointing in the right direction, focussed precisely on the source of the transmissions.

Steering the antenna and decoding the incoming signals required the building a new supercomputer with a plethora of high-bandwidth links to a constellation of high-orbit satellites, situated well away from the radio noise of the Earth. The hardware itself was more-or-less stock, if your pockets are deep enough - or at least those of your corporation or government - and getting the suitably modified satellites into orbit was a standard commercial operation these days.

It was the software was the real problem; this had to be custom-designed and built from the ground up. And this was my challenge to deliver. Just my luck, really.

Part 1 Part 3