The future of computing

For those of you unfortunate enough to have missed this morning's Inventions that Changed the World presented by my favorite over-patriotic journo-petrol head, Jeremy Clarkson, I'd like to relay some of the sentiments that I felt particularly affected by.

Who's heard of Tommy Flowers? No? I'm not surprised. I hadn't until this morning, or if I did his efforts weren't praised enough for me to appreciate them: He built the world's first code breaking computer - Colossus.

Now first of all, I'd like to say what a damn cool name for a computer Colossus is, I mean if a German informer had heard that the Brits had created a machine called Colossus that could break their Lorenz cypher in a matter of minutes I bet they'd have sheize'd their lederhosen! "Nein, eine Colossus komputer! Mein Furher, wir sind gefickt!"

Secondly, I'd like to say a personal thank you to Mr. Flowers for his tenacity to build the thing in the first place. Made out of nearly 2000 thermionic valves (vacuum tubes) it took only 10 months to build the first one (the British Army requested 8 more) and over £1000 of personal investment from Tommy himself. A post office worker by trade he used the same principle as the PO switchboard to run two messages simultaneously in perfect synchrony making code breaking a matter of minutes rather than months.

At this point I'd like to tip my hat to another fellow whose efforts we aught to recognize: Alan Turing. After being found to be a homosexual he was ostracised by the code-breaking community but his influence cannot be under emphasized. His Turing machine (or the principle behind the design) is the basis for pretty much all modern computing and his efforts not only on the bombe (a machine that sped up the breaking of an Enigma Machine encoded message) have undoubtedly saved lives. Alan is immortalised as a statue in Sackville Park, Manchester.

After waving the flag for British computing poineering it is unfortunate that our friend Mr. Clarkson has to head overseas to finish the story of the computer. The Americans took the idea of a programmable switching machine to new highs and build vast rooms filled with valves. Xerox, a New-York based company, were the poineers now and this group of mathematicians and hippies pushed forwards the potential for these machines. Originally set up as a company selling photographic equipment, mainly paper, they diversified in the 60s and started making paper copiers, the first named the Xerox 813 in 1963, before progressing onto computers.

In 1971 Intel created the first microprocessor: the 4004. Using two conducting layers separated by a semi-conductor, the state of which is controlled by an electronic current, the switching technology could be made incredibly small and efficient. With a clock speed of 740 kHz and 2250 transistors (which act as switches, the same as the valves in Tommy Flowers' machine) it was technically only a little bit more powerful than Colossus but it led the way for microprocessors becoming common fare.

Xerox built the first 'personal computer', called the Alto, in 1973 and it had features that are now commonplace: A cathode-ray screen, a GUI, mouse pointer and QWERTY keyboard. It was built around a processor based on Texas Instruments' 74181 chip, essentially a collection of logic gates but capable of 45454545.45 calculations per second (one every 22 nanoseconds). In 1981 the Xerox Star (with input from Apple) was the first commercially available system, but due to it's astronomical cost ($16,000) it still didn't sell well. By this time the internet was up and running (to a degree) and email was developed to take advatage of it.

Nowadays every electrical item in the home has a microprocessor that controls it's every function. Processor clock speeds topped out at just under 4GHz due to the heat emitted by the transistors changing state but faster machines have been built using a design that incooporates multiple processors on a single chip, which means that Moore's Law can continue unabated for the time being.

So, what lies ahead for computing? Mr. Clarkson's programme ends with a sarcastic run down of 80s science fiction suggestions: having conversations with computers, maids/butlers attending to our every desire, robo-pets that develop 'personalities'. Some of these are, on a basic level, already with us but the main thrust of his point, hidden and cryptic as it may be, lies in a comment he makes about half way through. On the subject of modern computers he points out obvious but not imediately comprehensible flaws in computer design: They will do exactly what we programme them to do, obediently and unquestioningly. The limit of computer intelligence is therefore limited to the height of human intelligence. This is, I believe, going to be the truth that computer scientists will have to mould to their whim if Artificial Intelligence is going to become a reality. In principle AI is therefore impossible because a computer is only as intelligent as it's programmer and no matter how many switches/transistors a computer is blessed with it will still only live in a world with electricity and humans to provide it with the 'food' and knowledge that it requires.

However, humans are, to a degree, also only a collection of switches (neurones) that relay electronic pulses to a central processing unit (the brain) which de-cyphers these pulses into what we see, hear, smell, feel and do. Is it so improbable that one day a computer/cyborg will be built that can take energy from it's surroundings (solar/wind power) with enough basic programming to learn how to make decisions based on input signals in various circumstances? I, for one, believe that it can be done and that the only thing holding us back is the lack of a unified international effort to combine the programming knowledge and learnt computer history of all AI projects across the globe. Whether I want this to happen is another matter entirely and for the time being I'd rather we didn't put too much funding into it. (Anyone seen iRobot?)

From another programme about the future, James May's Big Ideas, Mr. May explores the current level of bionics and robotics and is amazed to see how far we've come. One particularly impressive project is called Asimo, a Japanese robot that can walk, run and climb the stairs. Another project from the sme laboratory can recognise physical opbjects and will learn the name of a new or foreign object. It can only be a matter of time before the ability to interact with the physical environment and the ability to learn is combined to create what we could, I suppose, call AI. However, until I can't tell a human and a cyborg apart I will remain steadfastly asured that we are not threatened by a Hal that won't open the cargo-bay doors. That is, still, science-fiction.