The system results from Dr Tim Willis’ PhD research at the School of Informatics at Edinburgh University, originally aimed at computer users with motor disabilities.

Abbreviation Expansion (AE), which lies at the heart of Flexpansion, is an alternative to Word Prediction (WP), both of which aim to reduce the number of characters needed to type a text. As you type they offer a list of words matching your input so far, sorted alphabetically or by likelihood. Usually only the top 5 or so are displayed, so you want to reach the intended word as soon as possible for quick selection.

With WP you have to type every letter of the word from its beginning before you can take advantage of any unusual ones that set it apart. Since there are many shared sequences of opening letters in English (such as per-, under-, inter-), this can take a while. Also many words have early double letters (acc-, comm-, irr-, opp-).

With AE you can choose the most prominent letters in a word, leaving out vowels and double letters as you like, and using phonetic shortcuts, just like in text messaging or internet chat. For instance ‘f’ instead of ‘ph’ in ‘phone’, or ‘8’ instead of ‘aight’ in ‘straight’. Because it also takes into account context, often it will guess your desired word after only a couple of letters, so you can select it with ‘space’ or a full stop etc. and move onto the next. The start of a sentence is automatically capitalised too to save you time.

How many keystrokes and how much time you save overall depends on factors such as the vocabulary size you use, how many words you want displayed onscreen at once, and the type of device you are typing on. You can train the system easily on specialised language such as legal documents, business terminology or scientific fields.

You don’t need to memorise anything other than the fact you can leave out letters. With practice you will find new shortcuts, and as you go along, Flexpansion learns from you and improves its performance.