Keybord on the screen

For inputting characters only 8 of the 12 keys of the standard mobile phone keypad are used. In a general case it would be possible to use all 12 keys for input of characters. However, such approach will not be convenient because some keys shall also be used for other purposes, such as changing the case and input mode, inserting punctuation marks, etc.

Besides, as a rule, one key is designated for the “space” character, which is the most frequent character in all languages of the world. In case of simulation of a virtual keypad on the phone screen, in addition to the above-mentioned reasons, there is also another serious reason for minimizing the number of keys, used for input of characters, because 9 or 8 key will occupy significantly smaller space on the screen than 12 keys.

Therefore, it is necessary to input, with 9 or 8 keys, for example 26 characters (English alphabet) or 33 characters (Russian alphabet). If 8 keys are used, each key will account for 3-4 English letters or 4-5 Russian letters. If 9 keys are used, the achieved advantage will not be significant – 3 English letters per a key (with the exception of one key having 2 letters), or 4 Russian letters per a key (with the exception of one key having 5 letters).

During work on the “ARONETIS™” project, a decision was taken to use 8 keys, because in this case there will be more variants of their arrangement on the screen, while maintaining the unique association of the key images on the screen with the keys on the keypad, than in case when 9 keys are used. Besides, if each key on the screen in any time point represents only one character, then some key must be used for changing (shifting) the characters on the key images.

In the result of a number of experiments, a decision was taken to arrange the keys on the screen along its perimeter. This allowed to allocate less screen space to the key images and, accordingly, to leave more space on the screen for representation of the input text. In addition, such arrangement of the key images ensures the possibility to represent, if desired, more keys on the screen because there is still enough space along the perimeter of the screen, an in this case the input area will not change its form and screen coordinates.

Therefore, for input of characters, we have assigned the keys of the keypad located around the key “ 5”. Accordingly, for the mobile phone keypad, these keys are “1”-“ 4” and “6”-“ 9”. The numeric keypad of the computer keyboard is inverted as compared to the mobile phone keypad, therefore, in this case the used keys will be “7”-“ 9”, “ 4”, “ 6” and “1”-“ 3”. The key “ 5” is located in the center of the keypad, accordingly, this key was assigned for shifting the groups of characters on the key images on the screen.

Forming Groups of Alphabet Letters

As it was already mentioned above, the principle of splitting the alphabet into groups of letters must take into account the frequency of letters in the text of the specific language ( more details).

It is evident that, in order to ensure efficiency of text input, the most frequent letters shall be inputted with minimum number of keystrokes, preferably with one keystroke. We may state with confidence that in the overwhelming majority of languages of the world (if not in all languages) the most frequent character will be “space”.

Due to this fact, in the mobile phones, the space character is assigned a separate key in the bottom row of the keypad (“*” - in “Motorola” phones, “ 0” - in “Nokia” phones, «#» - in “SonyEricsson” phones). Accordingly, in the “ARONETIS™” the space character is also assigned a separate key in the bottom row of the keypad – depending on the model of the mobile phone, the user may assign any of the keys “*”, “ 0” and “#” to be used as the space character. Detailed information on assignment of the space character over the keys of mobile phones produced by different vendors is provided in the table.

There are several methods to split the letters of the alphabet of a specific language into groups. The simplest method is to sort the alphabet letters by their frequency and divide them into groups of 8 letters each (in accordance with the number of keys used for input). However, this will not be the most efficient method.

Let us assume that characters of the first group (the most frequent characters) are present on the screen from the beginning, and then, for input of the characters of the send group, it will be necessary to press the key “ 5” and, after change of the key images on the screen, to press the key corresponding for the desired character of the second group. After input of this character, in order to ensure input of a character form the first group, the program should automatically switch to the characters of the first (primary) group.

If it is necessary to enter a character from the third group, the user will have to press the “ 5” key twice, and then on the desired character. After that the program will again automatically switch to the group of the most frequent characters (the first group ). Therefore, if we assume that there will be at least four groups (for English language) or five groups (for the Russian language), it will not be very convenient for the user to reach the characters of the last groups. Also, for the user it will not be easy to remember to which group the desired character belongs. The user will have to look in sequence through all groups to find the desired character and, though the user will have to go to the last groups more rarely than to the 2nd or 3rd group, this will be uncomfortable. Besides, if we introduce additional characters (such as punctuation marks) to the keypad layouts, then we will have even more groups with all negative effects involved.

In the result of such considerations, we have developed the scheme of splitting characters into groups, which is now used in the “ARONETIS™”. All letters of the alphabet and some punctuation marks were split into the following three groups. The 8 letters, which are most frequent in the real texts, were assigned to the first group. The 8 letters, which are less frequent, were assigned to the second group. All remaining letters of the alphabet and the most frequent punctuation marks were assigned to the third group. At that, the characters of the first group are always entered with one keystroke because after entry of any symbol of any other group the system automatically switches to the first group. Entry of a character of the second group will require two keystrokes – the user will press the key “ 5” to switch to the second group and the key corresponding to the desired character.

The third group is very large (it may include up to 24 characters). Therefore, to enter a character belonging to the third group, four keystrokes will be necessary. When the user presses the key “ 5” twice, to switch to the third group, all characters of this group will appear along the perimeter of the screen, all key images being visually grouped by three. Each subgroup corresponds to a key and, when this key is pressed, only the characters of this sub-group will remain on the screen. Now the user will enter the desired character of the selected subgroup by one keystroke. This method of organization of input of the less frequent characters proved to be most convenient of all the methods considered. With the help of this method it will be possible to perform input in the languages, whose alphabet includes up to 40 symbols. If we add two more keys (such as “*” and “#”) then, without changing the principle of input, it will be possible to increase the number of characters in the keypad layout up to 50.

Distribution of Letters in a Group

The third idea, on which the “ARONETIS™” project is based, is that an alphabet letter shall be arranged on the screen of a mobile device in accordance with its position on the standard PC keyboard. This means that, if, for example, the letter “X”, on the standard PC keyboard, is located in the bottom left area relative to the center of the keyboard then, on the screen of a mobile phone, this letter shall be also located in the bottom left corner.

This condition will significantly simplify text input with the “ARONETIS™” system for the PC users. When searching for the desired character on the screen, such user will first intuitively search for the character where he/she has used to see this character (in particular, on the PC keyboard).

Unfortunately, this principle may not be applied for all alphabet letters simultaneously because the “QWERTY” design of the keyboard does not correspond to the input matrix used in the “ARONETIS™”. Therefore, we will have to reconcile with several exceptions for the layouts of some languages.

About layouts in details

Comparison of the Systems of Text Input

As it was already mentioned the “ ARONETIS™ ” system has an advantage in efficiency over the systems of “MultiTap” type. If we compare the data of tables 2 and 3, then the advantages of the “ ARONETIS™ ” method will become apaprent. So, to input a text in Russian language consisting of 100 characters, the user will have to perform 157 keystrokes with the “ ARONETIS™ ” method and 208 keystrokes with the “MultiTap” method. For English language, the efficiency of the “ ARONETIS™ ” method will be even higher – the number of keystrokes will be reduced by 65 or, respectively, by 30%.

It should be noted that, in practice, the advantage of using the “ ARONETIS™ ” method will be higher because in the above example we did not take into consideration the punctuation marks while input of them with the “MultiTap” method requires at least 5-6 keystrokes depending on the model of mobile phone.

However, it is not so easy to assess efficiency of the “ ARONETIS™ ” method relative to the predictive input methods (“Т9™”, “iTap™”). From a theoretical point of view, these systems solve the problem, which is different of that of the “ ARONETIS™ ”: the systems like “Т9™” eliminate ambiguity of text input with the keypad at the level of words, while the “ ARONETIS™ ” system eliminates ambiguity at the level of characters.

Besides, it should be noted that realization of the “Т9™”-like systems significantly depends on the model of model phone, namely on its memory and, accordingly, the stored vocabulary. Also, it is necessary to know frequency of specific words and the required number of keystrokes for each such word. The matter is in that, even if the vocabulary memory of a “Т9™” or “iTap™” system will include all words, the average number of keystrokes per one character will still be more than 1 because several words fall on a combination of keys. In case the entered word has less than 3-5 characters, the number of variants could make up a greater number than the number of characters in the word. In such cases, the actual number of keystrokes for entering the word will be more than two keystrokes per character.

Besides, if some word is absent in the stored vocabulary, the user will have to enter such word with the "MultiTap” method. At that, there may be quite many such words (personal names, abbreviations, technical terms, etc.). In such cases efficiency of input will be even lower than with “MultiTap” system because the user will have to spend additional efforts to determine that the desired word is absent in the vocabulary.

In the above-mentioned article by J. Arnott and M. Javed, the authors conclude that it is quite promising to eliminate ambiguity of input at the level of characters as compared to eliminating ambiguity at the level of words. It should be noted that system ы of predictive input were already known in the art when this article was written. The authors of the “Т9™” system also recite this article in the descriptions of their patents, which are based on the [2 -5] system. They speak in support of this system and try to refute the conclusions of by J. Arnott and M. Javed in practice.

However, correctness of this thesis by J. Arnott and M. Javed was proved by the “ARONETIS™" system where there is fundamental possibility (on condition of insignificant rework of the application) to use the touch-type method. None of the above-discussed systems will allow, even theoretically, to organize input of characters, using the keypad, with the touch-type method i.e. without looking either at the keys or at the screen. However, this is enabled with the “ARONETIS™" system.

We test the technologies experimentally. The text is the famous phrase used for SMS-typing contests:

"The razor-toothed piranhas of the genera Serrasalmus and Pygocentrus are the most ferocious freshwater fish in the world. In reality they seldom attack a human".

The results are following: ARONETIS takes 260 keypresses, Т 9™”- 264 and “MultiTap”- 340 ones.

About the results in more detail

Conclusion

The “ARONETIS™" system may be used in all applications where it is necessary to input text with the keypad (reduced keyboard), such as editing records in an electronic organizer, entering inquiries in pocket translators, entering passwords, etc. It is natural that, if vendors of mobile phones add support of the “ARONETIS™" system in their phones, then use of the “ARONETIS™" will be transparent for the application developers and will not require to increase size or other parameters of an application.

Use of the “ARONETIS™" will be particularly practical to organize input of strong passwords. In the future versions of applications, based on the “ARONETIS™" system, the user will be able to create own layouts, which will ensure easy input of any combinations of Unicode characters.