Journal Search Engine
Search Advanced Search Adode Reader(link)
Download PDF Export Citaion korean bibliography PMC previewer
ISSN : 1598-7248 (Print)
ISSN : 2234-6473 (Online)
Industrial Engineering & Management Systems Vol.16 No.2 pp.215-223

Optimized Layout of a Medium-Type Model for Korean Soft Keyboards in Smart Phones

Kuk Kim*
Department of Industrial and Management Engineering, Seokyeong University, Korea
Corresponding Author,
January 1, 2016 March 9, 2017 March 13, 2017


Medium-type Korean soft keyboards for smart phones are required currently and are considered an optimized keyboard type that combines a QWERTY-type keyboard, which suffers from small key size, and a 12-key keypad, which require a troublesome input method. In this paper, we will look at the design of medium-type Korean keyboards consisting of 3 rows. The models are classified as having one of the following layouts: (A) a lattice layout with a backspace, (B) a lattice layout with a backspace and a shift (auxiliary) key, and (C) a tilt layout. The models can also be classified as 6-column, 7-column, and 8-column models. The lattice layout of the 7-column model is developed representatively (as a prototype), in which 14 keys for 14 consonants and 6 keys for 10 vowels are arranged. An experiment is performed to obtain the difference in typing times between a QWERTY-type soft keyboard and the proposed model; the p-value of the results is 0.00114. Therefore, the developed model is significantly more effective than the QWERTY-type model. Digits and symbol characters are allocated to the shoulder positions and are accessed using the long-pressing input method.



    The keys on the QWERTY-type soft keyboards of smart phones are very small compared to the size of the average person’s finger, but on the other hand, the 12-key keypad is inconvenient as it has a troublesome input method. Kim and Yoo (2008) studied a new Korean computer- keyboard design with frequency and standard consideration. They redesigned an array of the QWERTY-type hard keyboard, which improves the finger load by calculating the position score. In reality, it is difficult to overturn the standard because of reverse learning and social cost, and there is the problem that a lot of the keyboard area is required for a soft keyboard. Kim (1992) studied a common approach to designing the Korean keyboard, and he tried to develop a 3-set type of Korean hard keyboard without the two-set problem of syllabication. However, the 3-set method is not desirable due to the large keyboard- area requirement and high complexity required for a soft keyboard. The National Institute of the Korean Language (2007) studied the arrangement of Korean computer keyboards and observed that the Korean keyboard uses 7 shift keys of twin consonants; thus, the number of shift keys needs to be reduced. The effort to eliminate the use of shift keys is meaningful, but it also has limitations on a soft keyboard.

    We want to study a medium-type Korean soft-keyboard design. Seminars concerning medium types are often held, such as the Standardization Committee of NXM Keyboard (2012a, 2012b). However, these seminars are centered around qualitative and self-directed arguments, and logical research papers are virtually impossible to find. This paper may be a starting point for such logical research papers.

    In general, smartphones are approximately 70 mm in width and 140 mm in height, with variations in the dimensions due to different makers and models. The keys of QWERTY keyboards are approximately 5 mm in width, which is much narrower than the keys of 12-key keypads as well as the fingering-width recommendation (19 mm) of human-factors engineering. Hence, touching errors during text input frequently occur due to finger interference. On the other hand, the number of basic assignments of alphabets of 12-key keypads is smaller than that of QWERTY keyboards, as the number of touches per text input is greater than that of QWERTY keyboards. Therefore, the optimal medium-type design for fast typing may exist as a combination of the 12-key keypad and QWERTY-type keyboard.

    It should be noted that this approach to medium-type keyboards is available due to the properties of Korean vowels (for example, the ‘ㅏㅏ’ form does not exist, so ㅑ=ㅏ+ㅏ is possible), i.e., the number of vowel keys can be reduced.

    In addition, on the touch screens of smart phones, the height of the soft keyboard and the height of the display window are mutually spatially exclusive. The 2-row model is inadequate because it is virtually unusable due to the width of the phone. The 4-row model merely adds an extra row of digits, which is advantageous for inputting digits. However, due to spatial exclusivity, the 3-row model is selected because the digit row can be replaced by the long-press method (see section 3.3).

    Medium-type Korean soft keyboards for smartphones are now available that have a larger key size than the QWERTY type and a less cumbersome tapping method than the 12-key keypad. We want to develop a medium- type design that is faster than existing standard keyboards.


    2.1.Qwerty Type

    Examples of QWERTY-type Korean soft keyboards of smart phones are shown in Figure 1. The layout of the 26 letters is arranged in the same way as the keyboard layout of Korean Industrial Standards (2007).

    The Korean alphabet consists of 24 letters, ㄱ, ㄴ, ㄷ, ㄹ, ㅁ, ㅂ, ㅅ, ㅇ, ㅈ, ㅊ, ㅋ, ㅌ, ㅍ, ㅎ, ㅏ, ㅑ, ㅓ, ㅕ, ㅗ, ㅛ, ㅜ, ㅠ, ㅡ, and ㅣ. The first 14 letters are consonants and the latter 10 letters are vowels. Although the Korean alphabet consists of 24 letters, not 26, the QWERTY-type Korean keyboard consist of 26 basic keys, which are the 24 letters and the 2 ligature forms, ㅐ(←ㅏㅣ) and ㅔ(←ㅓㅣ).

    The shift keys consist of 7 letters, which are ㄲ, ㄸ, ㅃ, ㅆ, ㅉ, ㅒ, and ㅖ, and the remaining shift keys are the same as the basic keys.

    The greatest property of the Korean script (Hangul) is the ‘compound syllable’, which consists of a first consonant( s), vowel(s) and final consonant(s). The final part may or may not exist. A syllable is a graphically formed ligature, for example, ㅎ+ㅏ+ㄴ = 한 (Romanized as H+A+N = ). Each part consists of one or more letters and forms a ligature itself as ㅘ (←ㅗㅏ) of vowels or ㄲ (←ㄱ+ㄱ), ㄳ (←ㄱ+ㅅ) of consonants. In modern Korean, ㅗ+ㅏ=ㅘ is available, but ㅗ+ㅓ is not available. The fixed rule for a compound syllable is shown in Table 1.

    Ligatures, syllables and texts are generated automatically on the Korean QWERTY keyboard as “ㅎ+ㅏ+ㄴ+ㄱ+ㅡ+ㄹ = 한글”. However, the 5 double consonants, ㄲ, ㄸ, ㅃ, ㅆ, and ㅉ, are needed to avoid conflicts between syllables. For example, a sequence “ㅇ+ㅣ+ㄹ+ㄱ+ㄱ+ㅣ” automatically generates “읽기” and not “일끼”. To generate “일끼”, ㄲ needs to be input as one letter, for example, “ㅇ+ㅣ+ㄹ+ㄲ+ㅣ”. Currently, the “ㄲ” key is the shift key of ㄱ, so this key is pressed sequentially after the “Shift key” on soft keyboards.

    According to the fixed rule of forming ligatures, there are no double vowels of the same type; for example, ㅏ+ㅏ cannot form a double vowel. Therefore, the multitapping of a vowel key can be defined as another vowel without causing any conflict in the input text.

    There are some reduced models, as shown in Figure 2, in which the consonant keys are the same and the vowel keys are reduced (compare with Figure 1).

    2.2.12-Key Keypad

    ITU-T E.161 (2001) defined the assignment of the basic 26 Latin letters to the 12-key telephone keypad.

    European Telecommunications Standards Institute (1996) and ISO/IEC 9995-8 (2009) also addressed this. Language-specific letters (e.g., ü, é, å, ä, ö) as well as other characters (e.g., ‘€’ or ‘@’) are not addressed.

    European Telecommunications Standards Institute (2007) specified the assignments of graphic characters (letters, digits and special letters, e.g., ü, é, å, ä, ö) to the standard 12-key telephone keypads provided in hardware form (e.g., as push-button keys) or software form (e.g., as soft keys on a visual display).

    In Korea, Korean Industrial Standards (2011) specified the arrangement of the Korean letters on 12-key keypads after much consideration over the past 30 years because of various designs and arguments. Uses for this mapping include the multi-tap text systems. Figure 3 (A) shows a visual marking, except for the 5 double consonants that are included in the corresponding single consonants.

    “ㅣ”, “ㆍ”, and “ㅡ” are strokes of vowels, such that ㅏ=ㅣ+ㆍ, ㅑ=ㅣ+ㆍ+ㆍ, etc. (“ㅣ” and “ㅡ” are themselves 2 of the 10 elementary vowels.) This is the most significant issue because “ㅣ”, “ㆍ”, and “ㅡ” are only strokes of letters, not letters themselves. In a figurative sense, in English this would look like A = / \ -, B = |)), etc., after defining strokes of letters such as /, \, -, |, and ‘space’).

    Korea Telecommunications Commission (2011), a Standard of the Korea Telecommunications Commission, defines supplements to the second and the third arrangement as shown in Figure 3 (B) and (C), where neither show the 5 double consonants visually.

    The first assignments are provided in both hardware form and software form, but the second and the third assignments are provided only in software form.

    The input method of the second arrangement, shown in Figure 3 (B), is multi-tapping and converting, and there are two conversion keys. The “획추가” key (which means ‘adding strokes’) is a conversion key, for example, ㄱ→ㅋ and ㅅ→ㅈ→ㅊ; the “쌍자음” key (which means ‘doubling’) is another conversion key for double consonants, for example, ㄱ→ㄲ and ㅅ→ㅆ.

    The unbalanced number of letters assigned to the keys and the complex fingering required for using two conversion keys are the biggest disadvantages; for example, the sequence to generate “ㅉ” is ㅅ+“획추가”+“쌍자음”.

    The input method of the third arrangement, shown in Figure 3 (C), is multi-tapping. The biggest disadvantages of this method are that the separator is located outside the keypad, and there is no redundant key on the keypad.

    Kim (2011) studied these problems and derived an optimal design (Figure 4) based on the analysis of the following three quantitative performance indices: stroke number, movement number, and moving distance (the smaller, the better). “Stroke number” is defined as the average number of strokes per Korean-alphabet input, “movement number” is defined as the average number of additional key movements per Korean-alphabet input, and “moving distance” is defined as the average movement distance (one unit = one key width) per Korean-alphabet input. These indices take into consideration the occurrence frequencies of Korean letters. Furthermore, the integrated index is defined as the product of these 3 indices. The performance indices of the models of Kim (2011) are shown in Table 2.


    3.1.Approaches of Medium-Type Soft Keyboard

    We will design a medium-type soft keyboard in this paper in which more alphabet letters are assigned than the basic letters in a 12-key keypad, with a larger key size than that of a QWERTY-type keyboard. The Standardization Committee of the NXM Keyboard (2012a, 2012b) presented various suggestions of medium models from individual inventors. However, these suggestions are not based on scientific studies and have only qualitative characteristics.

    Unlike the upper- and lower-case letters found in the Latin alphabet, there are no upper-case letters in Korean. In Korean, text can be formed without the need for shift keys. However, we note that Korean consonants are used as first consonants and final consonants of a syllable ligature.

    Korean vowels have a beneficial trait in that there is no doubling of the same vowel (no ‘ㅏㅏ’, compared with ‘aa’), and they also have fixed rules for ligature forms, as shown in Table 1.

    Due to this property, if at least 3 keys are used, all vowels can be input without any conflict: ㅏ, ㅓ=ㅏ+ㅏ, ㅑ=ㅏ+ㅏ+ㅏ, ㅕ=ㅏ+ㅏ+ㅏ+ㅏ, ㅗ, ㅜ=ㅗ+ㅗ, ㅛ= ㅗ+ㅗ+ㅗ, ㅠ=ㅗ+ㅗ+ㅗ+ㅗ, ㅣ, and ㅡ=ㅣ+ㅣ, considering the frequency. The tradeoffs are only the number of tappings and the number of vowel keys.

    Although the 3-stroke model of “· ㅡ ㅣ”, known as “Chon-ji-yin”, of a standard 12-key keypad is familiar to most Korean people, these symbols are only strokes of vowels, which is not adequate in terms of Korean script orthography. Therefore, we will assign the vowels themselves.

    It may be possible to construct models that use vowel keys, as shown in Table 3.

    Two ligatures, ㅐ (←ㅏㅣ) and ㅔ (←ㅓㅣ), are troublesome because they have the same status as ㅘ (←ㅗㅏ) graphically. They can be constructed using basic vowels according to the rules of orthography and are not essential key assignments.

    Practical designs require at least 14 consonant keys and 3 vowel keys; hence, the number of columns must be six or more.

    Each model is classified as having either a lattice layout or a tilt layout, as shown in Figure 5. The tilt layout loses one key in the middle row, but the center-to-center spacing distance of the two adjacent keys between the upper and the lower row is larger than in the lattice layout by the Pythagorean theorem (Kim, 2014).

    Lattice layouts are classified further as lattice layouts with a backspace and lattice layouts with a backspace and a shift (auxiliary) key.

    The order in the Korean alphabet is consonants first, then vowels. Therefore, the alphabetical-order arrangement naturally gives consonants the upper domain and vowels the lower domain.

    A Korean syllable only consists of a first consonant part on the left, and then a vowel part; hence, the left consonant domain and right vowel domain are familiar to Korean people.

    Our designs are not for a computer keyboard but for a smartphone soft keyboard, where finger sharing is not used but, rather, one or two thumbs are used. Hence, we arrange the letters in alphabetical order, i.e., ㄱㄴㄷ (similar to ABC) in a lattice layout, considering the do- mains of consonants and vowels. It is better to memorize the positions of letters, but the arrangement of a tilt layout will follow the Korean QWERTY-type keyboard. In addition, we consider the 6-column, 7-column, and 8-column models, as shown in Table 4. Digits 0-9 are assigned to shoulder positions over basic letters, which will be explained in section 3.3.

    The tilt layout has the merits of having compatible features with those of a QWERTY keyboard and a shift key and a backspace key with natural positioning. However, medium-type models cannot fully be converted to a QWERTY-type keyboard; hence, the lattice layout is better for medium-type models.

    In lattice layouts, the key size of a 7-column model is larger than that of an 8-column model. The number of taps on a 7-column model is less than that on a 6-column model. Therefore, a 7-column model is optimal for a Korean medium-type model for smartphones, should one have to choose between a 12-key keypad and QWERTY keyboard, as shown in Figure 6.

    We need an essential key to generate double consonants.

    • 1) Traditional shift-key method: as Shifter +ㄱ → ㄲ.

    • 2) Separator-key method: as Separator+ㄱ+ㄱ → ㄲ.

    • 3) Conversion-key method: as ㄱ+Converter → ㄲ.

    The converter is the same as the reverse shift key. The shifter is used via “see and select”, but the converter are forced to “infer and apply”. The key ⇧, shown in Table 4, can perform the role of “shift”, “separating”, or “converting” with appropriate marking. This key with this role in model A is positioned outside.

    A double consonant is only two instances of a simple consonant, for example, tt = t + t. Therefore, generally speaking, a separator is reasonable for Korean, which has no shifted alphabets. In addition, a separator can be applied to old Korean script that has a hetero type, such as ㅺ, of the first consonant. Another input method of double consonants is long pressing, which is presented in section 3.3.

    3.2.Function-Key Domain

    The function-key domain needs the following keys: spacebar, enter, Korean-English conversion, and symbolkeyboard conversion, among others. A system-set key and a voice-recognition key may also be necessary. An auxiliary (shift) key is needed in the function-key domain for Model A, but a backspace is in the right-hand corner of the third row of every model in Table 4. This issue of the function-key domain will be studied more in the future.

    3.3.Digits and Symbols and the Long-Pressing Method

    The standard arrangement and numbering for push buttons corresponding to the digits is as shown in Figure 7 (a) (ITU-T E.161). For a 2×5 array or a 5×2 array on a special telephone apparatus, the array is as shown in Figure 7 (b) or (c). Additionally, the phone number may be input from the row of numeric keys of a keyboard, as shown in Figure 7 (d). In this paper, only the array shown in Figure 7 (b) is available for models.

    Digits and symbols are assigned to shoulder positions over basic letters and are obtained by using the longpressing method. There are hidden symbols that should be analogized as characters of the visible shoulder characters.

    Another input method for double consonants is long pressing. Double consonants are assigned as {ㄲ 1} over ㄱ, {ㄸ 3} over ㄷ, {ㅃ 6} over ㅂ, {ㅆ 7} over ㅅ, and {ㅉ 9} over ㅈ. Though these 5 letters are hidden visually, their order of appearance is first. This longpressing method and the syllable-separator method are used together in this paper.

    Here, a prototype of Model A7 is developed based on the above approaches. Please refer to Figure 8 and Table 5, where the “한/영” mark is the Korean-English conversion key and the “↕구분” mark is a syllable separator, which can become a conversion key by dragging it up and down. The symbols in Table 5 are symbols on the ANSI (1988) keyboard, along with a middle dot and a Won-sign (₩, the currency symbol of Korea), that are frequently used in Korean.


    In past studies, no test results from medium models were obtained. In this study, first of all, a popular standard QWERTY-type Korean soft keyboard (Figure 1(d)) and the proposed medium-type model (A7) are tested in terms of typing times.

    Traditionally, a typing-proficiency certificate (TPC) rates the performance of a typist in the following two ways: 1) typing speed and 2) accuracy. The typing speed is, for example, the average words per minute (WPM) based on three 5-minute periods of typing. Characters per minute (CPM) may be used instead of WPM to Korean, and the accuracy (%) comprises how precisely the text sentences match, including punctuation with periods, commas, and spaces. Now, in the computer age, an accuracy of 100 percent can be achieved, reflecting the errorcorrection ability of a keyboard model. Further, the typing time is the reverse indicator of the typing speed: for the former, a lower value is better; for the latter, a higher value is better.

    In this paper, we use the typing time required to input a given text as the performance index. The devices of the experiment are the built-in QWERTY keyboard and a prototype of the proposed model (Figure 8) in smartphones. The text for the experiment, called “National Education Charter (Korea)”, consists of 393 Korean characters (962 Korean alphabets) and 120 periods, commas, and spaces. The subjects are 6 people with ages 21 (man), 21 (woman), 45 (woman), 47 (man), 57 (woman), and 59 (man). The text is presented to the subjects on printed paper. We ask the subjects to input the text according to their own personal preferences, such as typing with both thumbs or with one index finger. Subjects correct the errors that occur during the experiment using the backspace key or by touching the screen with a finger.

    The subjects are familiar with QWERTY type, so for the proposed design, we allow one practice period for each subject. Then, three experiments are conducted at a subject’s normal pace, and the median values are used as the data for each subject. Of the QWERTY type and the proposed type, the QWERTY type is used first to input the text by the subjects. We then ask the subjects to perform the test at a normal pace.

    3.5.The Results

    The results of the experiment are summarized in Table 6 and Figure 9.

    Statistical testing with the paired-sample t-test is conducted using Excel 2013 to examine the difference in efficiency between the QWERTY keyboard and the proposed model. The mean difference in typing times is 383.33 seconds, and the standard deviation is 164.09 seconds. The p-value of the test is 0.00114 by t-test; therefore, Model A7 is significantly more effective than the QWERTY-type model (see Table 7).


    Medium-type Korean soft keyboards with 3 rows for smart phones are designed in this paper. An auxiliary key is essential for the input of Korean text, whose role is as a shifter, a separator, or a converter (posterior shifter) specifically used for double consonants.

    Models are classified as having a (A) lattice layout with a backspace, (B) lattice layout with a backspace and an auxiliary key, and (C) tilt layout. Models are further classified as 6-column, 7-column, and 8-column models. All models have a domain of 14 consonants on the left side and a domain of 3-8 vowels on the right side to form the 24 letters in the Korean alphabet.

    Layout A of the 7-column model is designed representatively, in which 14 keys for consonants and 6 keys for vowels are arrayed. The proposed Model A7 is statistically more effective than the popular QWERTY-type model, according to the experiment on the difference in typing times, where the p-value of the results is 0.00114. The results of the experiment indicate that the proposed design is far superior, despite being less familiar to the participants of the experiment than the QWERTY-type design. We believe that this superiority is due to Fitts’ law from the perspective of Human Factors and Ergonomics.

    On the other hand, digits and symbol characters are allocated to the shoulder positions and are accessed using the long-pressing input method.

    Generally, the proposed model has the advantage of fast typing speed, but it has the disadvantage that it is not an accepted industrial standard.

    Studies on performance measures and validation data between medium-type soft keyboards should be performed in the future. In addition, assignment problems for function keys and symbol characters should be studied further.


    This research was supported by Seokyeong University in 2014.



    Examples of QWERTY-Type Korean Soft Keyboards.


    Examples of reduced models of QWERTY-type Korean soft keyboards.


    The first, second and third standard arrangements of Korean letters on 12-key keypads.


    Improved First Standard, Improved Second Standard and Proposed New Model of Kim (2011).


    The center-to-center distance.


    Effects of column models.


    Standard digit array in ITU-T E.161.


    Prototype of model A7.


    Typing time for keyboard layout.


    Korean syllable components to make a ligature

    Note: The Korean “ㅇ” has no sound at the first sound position and an [ng] sound at the final sound position.

    Performance Indices of the Models of Kim (2011)

    Models by vowels

    Models of medium-type soft keyboards (C: for consonants, V: for vowels)

    Hidden symbols of model A7

    Typing time to input the given text (seconds)

    t-Test table of paired samples of typing time

    *p <.05
    **p <.01
    ***p <.001


    1. (1988) ANSI-INCITS 154, ANSI Keyboard Layout , Available from:,
    2. (1996) Human Factors (HF): Assignment of alphabetic letters to digits on standard telephone keypad arrays, ETS 300 640, ETSI
    3. (2007) Human Factors (HF): User interfaces: Character repertoires, orderings and assignments to the 12-key telephone keypad (for European languages and other languages used in Europe), ES 202 130, ETSI
    4. (2009) Information technology -keyboard layouts for text and office systems-Part 8: Allocation of letters to the keys of a numeric keypad,
    5. (2001) ITU-T Recommendation E.161, Series E: Overall Network Operation, Telephone Service, Service Operation and Human Factors,
    6. Kim K (1992) A common approach to designing the Hangul code and keyboard , Computer Standards & Interfaces, Vol.14 (4) ; pp.297-325
    7. Kim K (2011) Standardization of Korean input keypad of mobile phones , Journal of Consumer Policy Studies Korea Consumer Agency, Vol.39 ; pp.75-87
    8. Kim K (2014) Symbol characters allocation of a QWERTY type keyboard design for smartphones , Industrial Engineering & Management Systems, Vol.13 (4) ; pp.408-413
    9. Kim K , Yoo Y K (2008) A new computer keyboard design for Korean alphabets with frequency and standard considerations , Journal of the Ergonomics Society of Korea, Vol.27 (3) ; pp.7-14
    10. (2011) Arrangement of Korean Character (Hangeul) on Telephones, KCS.KO-05.0046 ,
    11. (2007) Keyboard Layout for Information Processing, KS X 5002,
    12. (2011) Arrangement of Korean Character (Hangul) on Mobile Information Devices, KS X 5020,
    13. (2012a) Keypad layouts specified with three lines for inputting HanGeul, Standardization Forum of Korean Future Character Keys on the Information Devices(working document),
    14. (2012b) Keypad layouts specified with four lines for inputting HanGeul Standardization Forum of Korean Future Character Keys on the Information Devices (working document),
    15. (2007) The Study for the arrangement of Korean characters in computer keyboards ,