It is important to see how users perform using the T9 technologies because users do not expect that they will need training to use them. The more learning time they need for simple data entry, the more reluctant users will be to using newer technologies like T9, which might ultimately save them time. Even inputting data using regular keypads gets confusing when users try to enter special characters and punctuation. As the number of cellular phone users are growing, low performance times and high error rates will lead to frustration, unnecessary stress and will discourage novice and elderly from taking advantage of the newer technologies. Users must therefore, be able to enter text efficiently for storing phone numbers, retrieving text messages, and using newer services like email and web browsing.

Text entry evaluations are usually evaluated using two measures, speed and accuracy. Measuring speed can be done using a stopwatch and recording the time that users take to enter data using the two methods of data entry. Measuring accuracy is a little more difficult. Using "percent errors" to measure accuracy can be problematic unless data entry is constrained by forcing the subject to synchronize with the presented text [1]. To avoid this problem, it is therefore best to let the users type in the data and give them time to correct their entry. In this way, speed and accuracy can both be taken into account by measuring time until correct completion.

Subjective satisfaction is also important because if users do not like the newer technologies of text entry, they are less likely to buy a phone, which offers that technology. Users like to stick with things that are more familiar to them, that are available on older phones and that they already know how to use. Introducing new methods of data entry often causes apprehension and mistrust before the users actually use the technology. An empirical evaluation with users, is therefore, paramount to the viability of new methods of text data entry.

With no help or error checking available, it is vital to see how the users perform on text entry with minimal training. Certain issues should be identified and taken into consideration. For example, we must observe what the users do if they get stuck. They might simply keep trying until they are able to type in the correct letter. This becomes more difficult in T9 technologies, which works with a built-in dictionary and requires a single key press. It can be quite confusing at times because when the user looks at the screen they might not get the right feedback. For example, when the user tries to type "t", the phone screen might show a different letter on the same key as "t". Usually, not until the user types the last letter, does the screen guess and then show the correct word. It is interesting to observe if users get bewildered or frustrated because they are not getting the right feedback. They might choose not to look at the screen altogether while they type to prevent confusion. However this might not be such a bright idea, as Mackenzie points out "technology should evolve and bend to serve interaction" instead of "interaction confirming to technology" [2].

While comparing the two methods of text entry in mobile phones, it is important to realize how each of the methods work. The regular 12-key keypad requires less training but more keystrokes whereas the T9 method requires more training time but less keystrokes.

12-Key Keypad Method

Most of the cellular phones on the market use this as their text input method. Each key must be pressed one or more times by the user to specify a character. For example, to type the character 'A', the number key 2 must be pressed once. Typing 'B' requires pressing the number key 2 twice, and 'C' requires pressing it three times. When a character is placed in the same key as the previously entered character, most phones allow for a timeout period between 1 and 2 seconds. On some phones, the users can skip the timeout also known as "timeout kill" which allows them to directly enter the next character on the same key. Our experiment uses Nokia phones which includes both a 1.5 second timeout and the ability for a timeout kill using arrow keys so users can decide which strategy to use.

T9 method

This recent method, uses a built-in dictionary and adds knowledge to the system itself. It only requires one keystroke. For example, to enter "this", the user enters 8-4-4-7. T9 then uses the combination of letters and compares it to the word possibilities in its dictionary to then "guess" the intended word. This creates problems as many multiple words may have the same key sequence. T9 then guesses the most common word. Users can then press * to view the next possible word. The Nokia seventy one sixty is one of the phone manufacturers which uses T9 as one of its input methods.

The performances of the novice and expert users are very different. The novice using T9 for the first time will have to go through a training and practices session whereas expert users will already be adept. Expert users will already be familiar with the way T9 works. As Mackenzie also points out, "as expertise develops, users will invoke the timeout kill function". These are important issues to keep in mind and to keep the experiment consistent, we will only use subjects who are familiar with the 12-key keypad but have never before used T9 technology. This will enable us to see how quickly and in what manner the subjects can adapt to the less familiar technology.

A paper, which is very similar to our research, is Predicting Text Entry Speed on Mobile Phones by Scott Mackenzie and Panu Korhonen [3]. They present a model for predicting expert text entry rates for several input methods on a 12-key mobile phone keypad. In the paper, they found that for the traditional multi-press method, predicted expert rates vary from about 21 to 27 words per minute (wpm) whereas for the T9, expert rates varied from 41 to 46 wpm. Their analysis suggests "word-level disambiguation for English text with the traditional character layout on phone keypad is achievable with about 95 percent accuracy". They also mention that the overhead of interacting with not-so-perfect disambiguation of text entry degrades performance, but the cost is difficult to quantify because of the complex and varied strategies that the users can employ. We would like to see if we can confirm/validate his results in our experiments. Our subjects will be different, as will the tasks given to the users. We will only be testing 12-key keypad versus T9 instead of all the platforms of data entry used by them. Our subjects will consist mostly of college students who already know how to use 12-key keypad but are unfamiliar with T9 technology using a Nokia Seventy One Sixty. Although our subjects, tasks, and the platforms for our experiments are very different, we would still like to see if his results related to performance times and user satisfaction are validated by our experiment.

Determining which forms of data entry in cellular phones is more efficient, is a very important one. As the world is becoming fast-paced, people have no time to press a key a number of times before they are able to enter a single character. On the other hand, users also have little to no time for training and frustration increases if what they see on the screen is incompatible with what they are entering. In the time versus efficiency tradeoff, it will be interesting to see whether users prefer higher efficiency or less training time.

With cellular phones becoming the current trend in communication technology, research into methods of efficient text entry in mobile phones is justified. Previous research has been done in this area and a predictive model has been suggested. We will use this model to correlate our findings. Our ultimate goal is to make sure that future research and development in cellular phone technologies will consider the users efficiency in inputting data. Furthermore, we will analyze user satisfaction with text entry in the absence of help and error checking.