Flexible Keyboard for Everyone

Abstract

A user has hard time doing touch typing with a software keyboard. As these senses highly contribute to allow a user to recognize the layout of the keyboard, a user often feels disoriented to cause more mistypes once the user loses the senses with a software keyboard. We think that an appropriate key is typed in touch typing with a hardware keyboard. A user approximates layout of the keyboard through feeling of the finger tips, and relative distances between fingers. But hardware keyboard has a problem. This problem is a user can’t approximate layout of the keyboard. A current software keyboard provides no straight forward solution. One more fact to be discussed here is that the size of the keyboard matters for the better touch typing. This paper proposes “flexible keyboard for everyone” to tackle this problem. “Flexible keyboard for everyone” is a tailored software keyboard for every single user. “Flexible keyboard for everyone” learns a user’s habit with two major strategies and tries to help the user the better touch typing. This method allocates wider area for a weak key to pick up unintentional mistypes and increase the rate of appropriate types of the weak key to accomplish the better touch typing.

1 Introduction

Tablet devices have infiltrated our everyday life. One writes with a software keyboard quite often on a tablet device or smart phone, and this situation leads a problematic phenomenon. A user has hard time doing touch typing with a software keyboard. We think that an appropriate key is typed in touch typing with a hardware keyboard. A user approximates layout of the keyboard through feeling of the finger tips, and relative distances between fingers. But hardware keyboard has a problem. This problem is a user can’t approximate layout of the keyboard.

A user approximates layout of the keyboard through feeling of the finger tips, and relative distances between fingers. Therefore, the key input in the home position and the relative position of the user is recorded in advance.

This paper proposes “flexible keyboard for everyone” to tackle this problem. “Flexible keyboard for everyone” is a tailored software keyboard for every single user.

We can expect users to touch typing easier using Flexible Keyboard than a general keyboard.

2 Related Research

2.1 Study to Fit the Hand Shape

Hisano and Shiduki [2] suggests a “software keyboard easy to touch typing” on the touch panel. Hisano point out that touch typing is not easy because “the position of the key does not adapt to the shape of the hand of the user”. Hisano’s method relies on the vision of the user when matching the keyboard to “hand shape”, and the user can arbitrarily change the position of the key. In order to make typing easy, it is important to match with the shape of the hand, but I think that shape change relying on vision cannot be said to facilitate touch typing which does not rely on visual sense. So, in this research, we use a method to change the shape of keyboard which does not rely on visual sense.

2.2 Improve Input Speed

In the research by Endo and Go [1], when inputting with a software keyboard, “input speed reduction” is a problem due to “move line of sight”. Endo are approaching problem solving from both key shape and visual aid in order to improve the input speed. In addition, the method of Endo refers to the input position of characters and the number of inputs when matching the shape of the key to the user. In creating a keyboard, the input position can be said to be indispensable information, but considering the ease of striking each key, we think that creating a keyboard with the number of inputs will hurt ease of eating.

Because Endo’s method changes the shape of the key by moving the center of gravity of the key according to the number of inputs when matching with the user. Indeed, by moving the key’s center of gravity. However, with this method, there is a difference in easiness of striking with keys with a small number of inputs and many keys. Therefore, in this research, even with keys with different number of inputs, ease of striking Is difficult to differ.

2.3 Change Determination Position

The research of Karashima and Yanai [3] improving “input of performance” without changing “layout of QWERTY”. Karashima are doing the approach of changing the judgment position of the input without changing the shape of the keyboard.

The method of Karashima refers to the data of the actual input coordinates for the user who can perform the touch typing. Based on the input coordinates, the actual judgment position is created, and the keyboard is adapted.

It is also based on the input coordinates other than yourself that you can touch typing when creating the judgment position of the keyboard. In other words, it is a judgment position which is generally easy to strike, but I think that small hands or big people who do not match the shape of a general hand, for example, cannot be said to be easy to strike.

It is important that keyboards are easy for general people to enter. But I think that a software keyboard which is not a physical keyboard can make it easy to use persons with characteristic hand size. This is because keyboards that can be easily struck by anyone can be reproduced as they are adjusted from the software side because software is used to display them. Therefore, in this research it is important to match with each user We propose a possible software keyboard.

3 Method

3.1 How to Obtain Coordinate Data

What is important for creating a keyboard is the touched coordinates. If you use the touched coordinates, it becomes data dependent on the appearance frequency of characters. For example, when acquiring touched coordinates from a word such as “apple”, a keyboard biased toward “p” is created. An unfair keyboard is created in ease of striking from data in which characters with high and low occurrence frequency are mixed.

As a solution, touch panel coordinates are divided into blocks. Then, the coordinates that can be input are replaced with the block. When multiple identical characters are input to the same block, the number of inputs is 1. As a result, data with a dense character input frequency is approximated to sparse data.

Figure 1 shows inputs in areas partitioned into blocks and input coordinates on the touch panel. In the area divided into blocks, the input becomes 2, and the input becomes 4 on the touch panel. By dividing the touch panel into blocks, it is possible to make inputs with close coordinates as one input. It can be expected to approximate to sparse data regardless of the input frequency of characters.

Fig. 1.

Input of divided keyboard and touch panel

3.2 Step

In this method, the key is placed at the position recognized by the user. And “Flexible Keyboard” is a tailored software keyboard for every single user. For that purpose, the step will be explained first.

• 1. Read the home position of the user.

• 2. Read the relative position of the home position assumed by the user.

• 3. Create keyboard based on the data read by step (1) and (2)

• 4. The user inputs text with the keyboard created in step 3. (Count mistypes for each key)

• 5. Increase the size of key with many mistypes.

3.3 Processing of Input Data

There is creating keyboard based on data input by user. However, when making a keyboard from data input at a point, it is difficult for the user to input. Because the keyboard made with dots is distorted. Figure 2 is the keyboard (QWERTY) that the actual keyboard and the user think are desirable. Software keyboard must be a territory like this ideal keyboard. Therefore, we add work to make certain distance a single input as shown in the Fig. 3. By increasing the input coordinates, the distortion of the keyboard created is reduced. A keyboard that is easy to input is created.

Fig. 2.

Actual and ideal keyboard

Fig. 3.

Input after processing

3.4 Decision of Key Placement

Based on the data acquired in steps 1 and 2, arrange the keys in the divided coordinates. Combine all the data of A to Z on the divided coordinates. (Figure 4) However, when synthesizing on divided coordinates, there are cases where data overlap at the same coordinates. We solve this overlapping problem by weighting keys and comparing the weight values.

Fig. 4.

Combine input data

As shown in Fig. 5, the distance of the input center coordinates is converted to weight. Compare the given weights and determine the key placement as A to Z keys. Therefore, the keyboard reflects the relative position of the key assumed by the user.

Fig. 5.

Convert from distance to weight

3.5 To Correct of Keyboard

A user inputs the home position key and other keys to create a software keyboard. In this chapter, the user uses the created keyboard and corrects it based on the number of mistypes.

In this method, keys with more than a certain number of mistypes are increase size. (Figure 6) Increasing size of the key makes it easier to push the target key and reduces the number of mistypes.

Fig. 6.

Increase size of key

4 Results

In the experiment, we cooperated with 10 engineers by part time job. They can all touch typing.

Table 1 showed the comparison result with touch typing of the number of mistypes. The proposed Flexible keyboard has a lower average number of mistypes than the general Google keyboard. In addition, the number ratio of mistypes was 3: 6, Flexible and Google.

Table 1. Comparison of the number of mistypes

Table 2 showed that dispersion of Flexible is smaller than Google.

Table 2. Mistype of average and dispersion

5 Discussion

It said that Flexible Keyboard is ease of use for various people and Flexible Keyboard has few mistypes in touch typing. Therefore, users have easy time doing touch typing with a software keyboard than general keyboard.

However, only six people out of ten have achieved miss type reduction. Three people have few mistypes in the conventional keyboard, so it seems that some problems exist. For example, he is accustomed to the size of a conventional keyboard for the user.