Whenever I tell people that I used to be a lexicographer, working on the Oxford dictionaries, they usually ask how compilers decide which words to include in dictionaries. Perhaps it would be useful if I try to answer this FAQ.

Most modern dictionaries are based upon usage, so lexicographers need to find out how the language is actually used. When I worked at the Oxford University Press, this was done by means of a reading programme. A team of readers, mostly volunteers, ploughed their way through books, magazines and newspapers, looking for examples of new words. or interesting citations for old words.

Of course, readers had to have (or develop) a feeling for English, and what might be new words and phrases. Naturally we couldn't cover everything printed in English, so we tried to sample all the different kinds of writing: from literary classics to tabloid newspapers; from scientific journals to detective novels; from Nature and The Lancet to the New Musical Express.

The readers copied their finds on to slips of paper, six by four inches, with the illustrated word in the top left-hand corner so that these cards' could be filed alphabetically.

When I started working as a lowly editorial assistant (with a lowly annual salary of £600), one of my first jobs was to copy on to cards all the entries from a dictionary of cinematographic terms.

When the lexicographer came to deal with a particular word, he or she could find a number of cards containing quotations which illustrated the various ways in which the word was used. This basic evidence was supplemented with research in books in the dictionary department's own collections and in the invaluable resources of the Bodleian Library and other libraries, including the British Library.

The use of a reading programme was one of the best ideas of the people who started compiling the famous Oxford English Dictionary. They organised voluntary readers to comb publications in search of useful citations.

Nowadays, the gathering of evidence about the language is much easier because of computerised databases of material, giving lexicographers access to hundreds of examples for particular words.

When it was decided to update the OED, the OUP's dictionary department was set up on the first floor of a house in Walton Crescent, which was inhabited by an OUP printer, whose family used the rest of the house (with a young child often running up and down the stairs!).

Later, the department moved to a more palatial home at 37a St Giles, and eventually into the OUP's main building in Walton Street.

From the information supplied by such sources as the reading programme, lexicographers had to decide which words deserved inclusion in each particular dictionary. One was always aware of the limitations of space.

When I compiled the Oxford Intermediate Dictionary, I had room for only 1,200 words, so I had to choose the ones that I thought would be most useful to the readership of children aged between nine and 13.

These included words that had recently entered the language, like graffiti, print-out and punk. For larger dictionaries, a useful rule of thumb was that a word is worth considering for inclusion if you have five citations for it from a variety of sources.

People often ask if they can invent a word and get it into the dictionary but this will only happen if enough people decide to use it, preferably in print. George Orwell's novel 1984 was so influential that his coinages like doublethink and newspeak entered the language, along with a new sense of Big Brother.

When you decide to include a word in a particular dictionary, you then have to ascertain its pronunciation, etymology and the ways in which it is used. For example, if it should be labelled as slang' or colloquial'.

Compilers of the big OED also try to find the earliest possible occurrence of each word in a printed source. This process was illustrated in the recent TV series Balderdash and Piffle, although the work is done mainly by library researchers, not by celebrities!

Compiling dictionaries is often likened to painting the Forth Bridge. As soon as you have written a dictionary, it needs updating because the language changes so swiftly. A conservative estimate suggests that five new words enter English every day.

In the 1970s, when I was working on the Supplement to the OED, we had to deal with novel words and phrases like word processor, floppy disk and Internet because it was the start of the computer revolution.

The 1970s also popularised such new notions as animal liberation, Sloane Rangers, kneecapping, contraflows and tailbacks, Thatcherites, bottle banks and E numbers.

Labour exchanges became job centres; family allowances were replaced by child benefit; and the music centre superseded the radiogram.

Tony Augarde is the author of The Oxford Guide to Word Games (OUP, £14.99) and The Oxford A to Z of Word Games (OUP, £4.99).