Texting is one of the killer apps for mobile phones, particularly in Europe and Asia. Short-messaging service (SMS), which delivers chunks of text with as much as 160 characters to mobile devices, has become a potent tool for business, as well as a revenue stream for mobile carriers.
More than 50 billion SMS text messages were sent in the first quarter of the year, according to the international trade group the GSM Association. The current global monthly average - 16 billion messages - is up from 1 billion per month in early 1999. Many of these messages are one-way transmissions, sending information from a corporate or other wireless server to mobiles in the field. But many are generated by users, one key tap at a time.
The problem, of course, is that entering the text on the numerical keypad of a mobile device is laborious to the point of being ridiculous.
A lexicon of shortcuts has cropped up to solve it. The BBC's Web site, for example, offers a dictionary of common abbreviations, including a new crop of emoticons more appropriate for mobile phones than the old smiley faces devised for use in e-mail and PC-based chat. For example, the characters _/_/ suggest "drinks for two." (To see other SMS abbreviations, check out the lexicon at the BBC Web site: http://www.bbc.co.uk/so/cool/mobiles/texting_abbrev.shtml.)
Despite the help of abbreviations, however, predictive text-input software is becoming increasingly popular, developed with the goal of making texting easier on the fingers.
Take, for example, Tejic Communications' T9 software, available on 180 different models of mobile phones, including models from Nokia, Samsung, Ericsson and Sony.
The software works by pre-loading the phone with a dictionary of common words and names. As a user presses individual keys, the software calculates which words the user is most likely to want to say. An example drawn from a demo on T9's Web site shows that when the keys 6-3-3 are tapped, the phone initially generates the letters o-f-f, based on the fact that "o" is the most commonly used letter on the 6 key, and "f" is the most common letter to follow 'o.' But then a user taps the 8 key after that, the software knows they're choosing between the letters t, u and v, and determines they're probably typing the words "meet" or "meeting" instead of the word "office." Then, it changes the previous letters in the word to match with the new course.
Using software is certainly easier than going it alone, but Howard Gutowitz, CEO of the 14-month-old startup Eatoni Ergonomics, thinks this system is still too confusing. His company, named after the six most commonly used letters in the English language, has devised new types of predictive-text software that rely on likelihood of occurrence, while allowing the user to combine, or chord, the keys.
"Chording has a bad name in the industry, because supposedly no one will learn it," Gutowitz says, but adds that he believes that people will eventually learn to combine two keystrokes on a mobile-device keypad, just as they use a shift key on a standard keyboard to generate capital letters. But it would cross the line to use three keys, as you would on a keyboard with the CTRL-ALT-DELETE command necessary to reboot Windows machines.
Gutowitz, a former mathematics professor, came up with the notion to develop predictive-text software several years ago, while his girlfriend was working as an operator at France Telecom, transcribing people's voice messages to text, so it could be sent out to pagers. Most messages fell into two categories: kids and teens contacting their friends, and adults making appointments to see someone with whom they were having an affair. Noticing the predictability of the messages, Gutowitz began developing systems that could anticipate the next letter or even the next word, based on the likelihood of their occurrence.
In Eatoni's Letterwise system, for example, tapping a number key delivers the most likely letter on that keypad. If it's
