USER INTERFACE – KEYBOARD AND KEYSHORTCUTS

Grzegorz Sowa, Zbigniew Filutowicz

IT Institute, Academy of Management, Lodz, Poland

(gsowa, zfilutowicz)@gmail.com

Abstract
This article presents interface-related problems arising from a traditional
keyboard. Some alternative suggestions for keyboard layout are discussed.
Several ways of analyzing interfaces are introduced: the keystroke level model,
the human processor model and cognitive walkthroughs.
Key words: user interface, keyboard, keyshortcut, layout.

1 Introduction

The ordinary keyboard has always been, and remains, the main interface
between a human and a computer. For this reason, the analysis of optimal
keyboard usage, which has been carried out for many years, is still important.
Despite firmly established the QWERTY standard, new layouts are con-
stantly appearing and being tested, and it is possible that new generations of
users may adopt some of these designs.
Graphic interfaces for operating systems and applications have evolved
gradually from symbol systems and are still closely connected to them. The
use of keyboard shortcuts to call up functions is still widespread and related to
the layout of the keyboard.

2 The history of the keyboard

2.1 Discovery

The modern computer keyboard is a direct descendent of the mechanical

typewriter. One of the first the typewriter makers, C.L. Sholes, also designed
the keyboard layout which is still in use today. At first, he used an alphabetic
layout, but it transpired that letter combinations commonly encountered
in English, such as „th‟ caused the thumbs to block the levers in the model
of typewriter then in use. The study of common letter pairs in English helped

149
 Sowa G., Filutowicz Z.

to minimalise this fault. Thus, the design of the keyboard layout was more
carefully thought out than is often believed. It was also patented in 1878.

2.2 Dvorak’s Keyboard

The most significant competitor to the standard QWERTY layout appeared

a long time ago in 1932 [12]. The essence of Dvorak‟s design is the position-
ing of the most commonly used letters in the home row, vowels on the left and
consonants on the right (AOEUI-DHTNS); this enables the fingers of both
hands to be used in quick succession when writing. Similarly, the next most
commonly used letters are in the upper row – which is easier to reach than the
lower row.
Careful research [2] has confirmed that transferring the writing process
from one hand to another is much faster than striking the keys with one hand.
This is backed up by the MHP (Motor Human Processor) theory. Salthouse
[13] expressed it concisely: „A character on the same hand cannot be initiated
with the cognitive operator until the motor processor execution of the previ-
ous character is complete‟.
Despite these theoretical advantages, Dvorak‟s layout has not caught on,
a textbook example of an invention that was made too late [3]. The difficulty
of remaking mechanical typewriters is no longer relevant, but the QWERTY
standard is further entrenched by keyboard shortcuts which use the function
buttons F1=F12, and the special keys Ins, Del, Home, End, PageUp, Page-
Down, PrtSc, ScrLk, Pause, Break, often in combination with each other or
with letters.
For practical reasons (the limited size of computers and finger span),
a keyboard cannot have too many keys. This limitation applies particularly to
laptops (not to mention mobile devices). The most important solution to this
are numbers (keys UIOJKLM) and arrows (WASD) positioned on letter keys
– conveniently placed on the QWERTY keyboard and entirely randomly in
the Dvorak layout. Similarly, the most popular editing shortcuts: Undo, Cut,
Copy, Paste are opened through the neighbouring keys Ctrl+Z, X, C, V – but
only in the QWERTY layout.

2.3 Variations on the Dvorak keyboard

Despite the lack of market success, the Dvorak keyboard is being devel-
oped. One of the projects involved adapting it for people who write in a suit-
able way: only with their right or left hands. Apart from disabled users, this
makes sense for those whose other hand has to be occupied with something
else (the mouse, for example). This idea, however, does not appear to have
a future – users have become used to the alphabetic layout on mobile phones.

150
 User Interface – Keyboard and Keyshortcuts

However, an interesting idea is R. Kaufman‟s suggestion [6] – a keyboard

adapted for programmers working in popular languages (Pascal, C, Java,
XML). The layout of the letters (in Dvorak‟s standard) is not important here
overall, as what is more important is the positioning of all the brackets and
special signs in a logical order in the upper row: (([{}(=*)+]!#), as well as the
unique number layout – separating odd and even numbers (7531902468). This
kind of modification of the standard keyboard appears to be realistic and use-
ful. Whilst it is true that entering the source code is not the most difficult as-
pect of programming, it appears in copious amounts, so efficiency is signifi-
cant here.

2.4 Programming language instructions on the keyboard

The simplicity of Basic inspired an idea which involved putting all Basic
instructions on the keyboard, and this was put into practice on the Sinclair ZX
Spectrum [14] computer.
Some examples: letter R: RUN, Y: RETURN, U: IF, D: DIM, F: FOR, G:
GOTO, N: NEXT and so on.
Unfortunately, this was the only way for instructions to appear in the code
– beginner users found this very problematic, and programming in Basic on
this computer is not remembered fondly.
The obvious and most significant drawback was the practical impossibility
of developing the programming language based on this concept.

2.5 Research into typing

The comparison of typing speeds has a long history. Extensive research

has allowed several useful conclusions to be drawn [2]. Alongside the afore-
mentioned advantage of alternate usage of hands, it has been established that
we type words rather than letters: reading the word or at least the syllable
before typing increases the typing speed considerably. The mental processes
are carried out simultaneously here, and the time is radically reduced. The
results of this research inspired similar treatment of the computer interface.

2.5 The significance of the keyboard

In today‟s world, efficient typing on a keyboard is, paradoxically, becom-

ing less relevant as an issue. Since a huge number of people use a keyboard
and type their own texts, with better or poorer skills, the efficiency of profes-
sional typists is not very significant in the larger scheme of things. Keyboards
for professionals should certainly be ergonomic, but the average user values

151
 Sowa G., Filutowicz Z.

above all the standard layout and the quality of the keys. The clarity of the
interface is much more important than productivity.

3 Models of use of the key interface

3.1 Modeling aims

The most obvious task when analyzing the interface is to compare existing
interfaces. A precise evaluation of the applications currently in operation re-
quires a methodical approach, general impressions are not sufficient. This is
significant for the process of choosing between many applications on offer.
Precise data from the analysis can be placed alongside the description in
words.
Another case is to design the interface for a new application. There is no
sense waiting to complete it, one should evaluate a model. Other tools are
important – one should simulate something non existing yet [5, 10].

3.2 The KMS model

One of the first and best known tools for analyzing screen interfaces is the
KLM-GOMS model [1]. The development of the Keystroke Level Model and
Goals, Operators, Methods, Selection rules shortcuts sums up its significance
clearly.
Basic interface operations have been defined as [7]:
- Keying – pressing and releasing a keyboard key (0.2 seconds)
- Pointing – pointing to an object on the screen with the mouse (1.1 s)
- Homing – moving the hand from the mouse to the keyboard (0.4 s)
- Mental preparing - thinking (1,2 s)
- Responding – waiting for the system to respond
Drawing time (or moving an object) using the mouse (the theory behind
this has existed since 1954! - [4]) was then added to these basic operations,
but the essence of this model has remained the same. The exact operation
times are of course approximate, but can serve as initial estimates in the
model evaluation of an interface. It is extremely important to note that an
interface does not need to exist for such an analysis to take place – a suffi-
ciently detailed design is enough.
9 standard steps are suggested when evaluating interface [7]:
1. Choose one or more representative task scenarios.
2. Have the design specified to the point that keystroke-level actions can be
listed for the specific task scenarios.
3. For each task scenario, figure out the best way to do the task, or the way
that you assume users will do it.

152
 User Interface – Keyboard and Keyshortcuts

4. List the keystroke-level actions and the corresponding physical operators

involved in doing the task.
5. If necessary, include operators in case the user must wait for the system to
respond
6. Insert mental operators in case the user has to stop and think.
7. Look up the standard execution time for each operator.
8. Add up the execution times for the operators.
9. The total of the operator times is the estimated time to complete the task.

The analysis of contemporary interfaces yields one basic conclusion. The

time taken to press a keyboard key is considerably shorter than the time taken
to move the cursor and press a screen button using the mouse. However, be-
ginner users in particular cannot be expected to remember numerous com-
mands and keyboard shortcuts. The theoretical advantage of speed when us-
ing only the keyboard is obscured by the time needed for thought – for re-
membering or reading the instructions in order to retrieve the right command
or shortcut.
For the time being, the best solution is a balance between entering com-
mands and shortcuts on the keyboard and the graphic interface based on
a scroll-down menu. Unfortunately, a large amount of shortcuts makes the
use of the keyboard more difficult, as unwanted options are called up acciden-
tally remarkably often.
The old standards are being replaced by new ones, which are developing in
proportions that are hard to grasp. The Word 2007 word processor currently in
use suggests a few hundred shortcuts to actions, which are of course also
available on the menu. There is no standardization here, different applications
use their own solutions, and there is little noticeable continuous development.
We remember the Pascal word processor from our own work experience days,
where the method of selecting and moving a block of text was in no way
reminiscent of the current popular method Ctrl C, Ctrl-V.
For the experienced user, progress is doubtful. In the Lotus 123 spread-
sheet, entering STD was enough to call up the function of calculating standard
deviation, while currently in the Polish version of the Excel spreadsheet, this
function is called ODCH.STAND.POPUL, and the only reasonable way of
calling it up is to choose this name from the scroll-down list.

3.3 Human processor model

Even everyday experience demonstrates that we can carry out many men-
tal and manual tasks simultaneously: watching the road in front of the car and
carrying on a conversation, typing on a laptop keyboard and watching the
screen, etc. Therefore models based on precise identification of a sequence of

153
 Sowa G., Filutowicz Z.

tasks and adding the times taken together are imperfect. Multimedia and the
multifunctionality of many contemporary tools complicates these matters even
more.
One of the solutions (to this) is Model Human Processor methodology [9]
– the development of GOMS based on the assumption that perceptual, cogni-
tive and motor tasks can be carried out simultaneously. Critical path dia-
grams, known from operational research, are used here to represent the rela-
tionships between tasks.

3.4 Cognitive Walkthrough

The methods mentioned above focus on measuring the time taken to carry
out a task by the user. This misses the main problem. The interfaces of pro-
fessional applications are practiced on for a long time and users become adept
at operating even flawed ones after some time: overall the functionality of an
application is more important than the quality of its interface.
In practice, there is often a different problem: will the user who sees the
interface for the first time understand what he or she needs to do without be-
coming discouraged before the task has been completed? It is estimated that
even the majority of internet shop transactions are abandoned because poten-
tial clients do not understand the procedure. Therefore, research into situa-
tions where the client does not know the structure of the task at the beginning
(the number of tasks or their hierarchy) is very significant. The structure is
discovered by the user during the use.
One of the research methodologies examining these kinds of users is Poul-
son‟s Cognitive Walkthrough [11].
The evaluation of the proposed interface consists of four elements:
1 Representation of the active aim of the user
2 Description on the button
3 Localisation and the picture of the button
4 Action taken after pressing
A very wide range of imperfections can interfere with the action the user
takes. The label may be ambiguous, there may be several similar buttons ac-
tive on the screen, etc.

4 The interface and the law

The interface cannot be hidden. The best companies‟ achievements are

available for all to see and imitate.
An important precedent here is a well-known conflict in the 1990‟s be-
tween Lotus and Borland about the menu in their competing spreadsheets [8].

154
 User Interface – Keyboard and Keyshortcuts

The beginning of the court summons explains the problem: “This appeal re-
quires us to decide whether a computer menu command hierarchy is copy-
rightable subject matter. (…) Borland included in its Quattro and Quattro Pro
version 1.0 programs “a virtually identical copy of the entire 1-2-3 menu
tree.”(…). In so doing, Borland did not copy any of Lotus‟s underlying com-
puter code; it copied only the words and structure of Lotus‟s menu command
hierarchy‟.
In the end, the judge ruled that the structure of a menu cannot be the sub-
ject of a patent.

5 Conclusions

An application‟s interface is often the deciding factor for its level of suc-
cess. For the user, the interface is the application.
The interface evolves very slowly, in contrast to the software it serves. In
contemporary interfaces, very old solutions exist alongside new ones. The
traditional keyboard layout, keyboard shortcuts, the concept of calling up
menu functions through selected letters – all of this was devised a long time
ago. Therefore, the design and evaluation tools which were used with old
interfaces are still useful now. Currently, the analysis of interfaces focuses on
simultaneous operations. In practice, however, how easy an interface is to
understand is more important than how fast it is to use.

References

1. Bonnie E. J., Kieras D. E., Using GOMS for User Interface Design and Evalua-
tion: Which Technique?, 1996, the Association for Computing Machinery, Inc,
10 June 1996, ftp://ftp.eecs.umich.edu/people/kieras/GOMS/Which-GOMS.pdf
2. Bonnie E. J., 1996, TYPIST: A theory of performance in skilled typing, Human–
Computer Interaction, 11, 321-355. http://cogtool.hcii.cs.cmu.edu/trac/export/
3275/docs/papers/TYPIST.pdf
3. David P. A., May, 1985, Clio and the Economics of QWERTY The American
Economic Review, Vol. 75, No. 2, Papers and Proceedings of the Ninety-
Seventh Annual Meeting of the American Economic Association., pp. 332-337.
http://www.econ.ucsb.edu/~tedb/Courses/Ec100C/DavidQwerty.pdf
4. Fitts P. M., 1954, 6 (June), The information capacity of the human motor system
in controlling the amplitude of movement, Journal of Experimental Psychology
47, pp. 381-391. Reprinted in Journal of Experimental Psychology: General,
1992, 121, 3, pp. 262–269,. http://eidetic.ai.ru.nl/egon/education/IntroCE/ litera-
ture/classics/Fitts54 Information_capacity_motor_system.pdf
5. Halverson, T., Hornof A. J., 2007, A Minimal Model for Predicting Visual
Search in Human-Computer Interaction, CHI 2007 Proceedings, Gaze & Eye

155
 Sowa G., Filutowicz Z.

Tracking April 28-May 3, 2007, San Jose, CA, USA www.citeserx.ist.psu.edu/

viewdoc/ download?doi=10.1.1.109.5822.pdf
6. Kaufman R., Programmer Dvorak Keyboard Layout, http://www.kaufmann.no/
roland/dvorak/
7. Kieras, D., 2001, Using the Keystroke-Level Model to Estimate Execution Times,
University of Michigan, www.pitt.edu/~cmlewis/KSM.pdf
8. Lotus Development Corp. v. Borland International Inc. U.S. Court of Appeals,
First Circuit March 9, 1995 http://digital-law-online.info/cases/
34PQ2D1014.htm
9. Lui, Yili; Feyen, Robert; and Tsimhoni, Omer., 2006, Queueing Network-Model
Human Processor(QN-MHP): A Computational Architecture for Multitask Per-
formance in Human-Machine Systems. ACM Transactions on Computer-Human
Interaction. Volume 13, Number 1, March 2006, pages 37-70,
www.citeseerx.ist.psu.edu/viewdoc/download?doi=10.pdf
10. Miller C. S., March 14, 2003, Relating Theory to Actual Results, Computer
Science and Human-Computer Interaction.
11. Poulson, P.G., Lewis, C., Rieman, J., and Wharton, C. (1992). Cognitive walk-
throughs: A method for theory- based evaluation of user interfaces, International
Journal of Man-Machine Studies 36, 741-773.
12. Rehr, D. The QWERTY Connection, http://home.earthlink.net/~dcrehr/
whyqwert.html
13. Salthouse T.A., 1986, Perceptual, Cognitive and Motoric Aspects of Transcrip-
tion Typing, Psychological Bulletin Vol. 99 No. 3, http://portnov.com/keyboard/
303.pdf
14. Spectrum Emulator - Basic Usage, www.guybrush.demon.co.uk
/spectrum/docs/Basic.htm

156