Beyond all this, changing from analogue to digital form gives cyberspace an extraordinary, even unique, characteristic: it transcends familiar physical limits, and this lack of familiar physical limits lies at the heart of the perplexing opportunities of cyberspace.
Much of the world we are used to is governed, in one way or another, by physical limits. Cars, no matter how expensive, can only go so fast. People, no matter how talented, can only work for 24 hours in a day (at most). Computers and digital communication networks do away with many such limits; in particular, they make it possible to copy, manipulate and transmit digital information essentially for nothing. This ability has complex and unpredictable consequences. Here are some examples.
An audio cassette tape wears out; and if it is re-recorded on to a new tape, the new recording is never quite the same as the original. In the same way, a photocopy is not exactly the same as the original.
In contrast, digital information can be reproduced perfectly, time after time. There is no mystery to this: if you were given a list of numbers, even written in bad handwriting, you could transcribe them accurately, and someone else could copy your work, and so on for ever. The recording has been, up to a point, set free from the medium on which it is stored.
Even if some numbers do get lost in transit (perhaps the bad handwriting became illegible in one or two places) there are ways to correct for that. For example, you could store everything twice, and reckon that the chances of an error happening twice in the same place would be low. Delightfully cunning ways have been developed to correct errors in digital documents with much less overhead than storing them twice. Your mobile phone is doing this every time you use it.
There is an irony here, because while digital information can indeed be reproduced perfectly, the physical devices on which it is usually stored are much less robust than the traditional analogue storage mechanism, writing on paper or papyrus. We can still read documents written hundreds, or sometimes thousands, of years ago. But a floppy disk deteriorates in a few years, and even if kept carefully there might be no floppy disk drives available in a hundred years, let alone a thousand. The more sophisticated the medium (hard disk, CD) the more sophisticated the device needed to read it, and the less likely such devices are to be available in a hundred years. Instead, better, faster, smaller devices will be in use; but that doesn't help with a hundred-year-old CD, even if the CD has not deteriorated physically, which no one knows for sure.
So digital information is simultaneously more robust and more fragile than analogue information.
A physical document can be photocopied in its exact form, and sent to others. In contrast an electronic document can be edited, reordered, reformatted (for example, changing the page layout or font), checked for errors of spelling or grammar, indexed, sent to others by email, published on the web, and so on. All of this is much easier because the information is stored digitally.
All sorts of digital information can be manipulated, not just text. Altering a physical photograph is difficult, and relatively easy to detect. Altering an electronic photograph, though, is just a matter of changing some of the numbers. Manipulation is so easy that it is hard to know what is visual truth. Another example is the way in which someone can sing into a microphone, a computer analyses the recorded sound, deduce that it is out of tune and then correct it before putting it out over the speakers. Indeed, one can already go further and have the singer's voice sound like someone famous, the voice transformed into the characteristic vocal harmonic spectrum of that famous person.
Information in analogue form more or less has to be transmitted all at once. But the Internet transmits digital information in small chunks called packets. Your email message (or whatever) is chopped into packets and fed into the Internet; sooner or later most of them will arrive at the destination; the destination will ask the sender to retransmit any packets that got lost; and finally it will reconstruct your original message. Nowadays, this happens for your phone conversations too!
All of this is very different from the analogue way of working. The old phone networks would dedicate a particular piece of wire to a single conversation, even during the parts of your conversation when neither of you were saying anything. The ability to chop everything into packets and then just process the packets individually in any order makes it possible to carry huge numbers of different messages, conversations, stock trades, web accesses, or whatever, on a single wire
It turns out that the cost of transmitting a packet from A to B no longer depends much on where A and B are. That is, in reality it is just about as cheap to send a packet from London to Hong Kong as from London to Hatfield. This 'death of distance' has yet to manifest itself in local rate phone calls direct to Hong Kong, but that is now due more to commercial and political factors than technical ones.
One of the litmus tests of whether something is in cyberspace is whether its location counts. If you were to ask, for example, 'Where is this digital object that I am using?', the answer would often come 'I don't know and it doesn't matter.' In using the web, you look at a page of information delivered to your computer from the USA, click on a link that implies the next information page comes from a non-profit company in India, then another that is in Australia and then . . . After a few steps the location ceases to matter. Information is digitized but its geographical physical location is not important. Even if one knew, the location could be changed very rapidly to another digital 'space' on earth, or even stored in a memory on a satellite. Just as the music recording was set free from the medium in which it was stored, so the digital object is set free from location. There must be some storage medium; the data must be somewhere. But what and where generally don't matter.
Cryptography concerns the 'coding' of data, so that it cannot be read by unauthorized people. Codes have been around for millennia, but computers have driven recent (post -1950) developments in so-called strong cryptographic techniques. These techniques make it possible to take any digital information (picture, sound, money, it doesn't make any difference) and encode it in such a way that no one who does not have the key to the code can understand it.
Strong code techniques raise important questions about privacy and security; they will have a role in cyberspace, but what that role should be is still under debate.
We are used to thinking of money, for example, as having some physical form that is very difficult to copy; now money takes electronic form and banks take enormous care not to duplicate it by accident. We are used to thinking of a book as something with physical form, consisting of a certain number of original, identical copies; now, an electronic document may be altered on a daily basis by its author, and may be read or copied at any moment.
Even the physical limits that computers do have are being driven back at an astonishing pace. In 1980 a personal computer with 64,000 bytes of memory was considered vastly luxurious. Now it is common to find personal computers with 60,000,000 bytes of memory. A now rather hackneyed, but still useful, analogy is this: imagine if, over the same period, your car went 1,000 times as far on a gallon of petrol, or your house was 1,000 times as big for the same money.
In practice, the principal limitation on the scale and sophistication of computer systems is no longer their speed or memory capacity, but our own inability to imagine, design, and build the software that animates them. The complexity of these software monsters is daunting, and tales of commercial software disasters are commonplace. We cannot blame the medium, as a carpenter or a bridge designer can. Instead, we are face to face with our own intellectual shortcomings - not a bad thing, perhaps.