• No results found

A new method for content selection on a mobile device

N/A
N/A
Protected

Academic year: 2021

Share "A new method for content selection on a mobile device"

Copied!
125
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

A new method for content selection on a mobile

device

JJ Esterhuizen

orcid.org/0000-0003-2238-049X

Dissertation submitted in fulfilment of the requirements for the

degree

Master of Engineering in Computer and Electronic

Engineering

at the North-West University

Supervisor:

Dr JC Vosloo

Examination:

November 2019

(2)

A new method for content selection on a mobile device i

ACKNOWLEDGEMENTS

I would like to express my gratitude towards:

1. The Centre for Research and Continued Engineering Development (CRCED) for giving me the opportunity to complete the degree and continue my academic career in the engineering field.

2. Enermanage (Pty) Ltd. and ETA Operations (Pty) Ltd. who provided the funding needed for the research that was conducted.

3. The peer-reviewers for expert feedback and constructive criticism.

4. Dr JC Vosloo, my promoter and Mr Jaco Prinsloo, my technical mentor during the study. They provided critical support and input that elevated the quality of research conducted.

5. All colleagues who supported my research by providing informal feedback to improve various aspects of the study.

6. Megan Lowes for proofreading and editing of the dissertation, ensuring the language and format are of the highest quality.

7. My parents, Jeff and Riani Esterhuizen, and other family members, who provided support and motivation throughout the study.

(3)

A new method for content selection on a mobile device ii

ABSTRACT

Title: A new method for content selection on a mobile device

Author: Jacobus Johannes Esterhuizen

Promoter: Dr JC Vosloo

Keywords: Mobile productivity; Document writing; Text selection; Mobile devices;

Gesture usage; User experience; Error rate

In this modern, fast-paced society people must be able to work while away from their desks. The rise and adoption of mobile devices have improved productivity for many people who can now start or continue work which was traditionally limited to a computer. Many still choose to limit these tasks to short communications or reading emails, due to a perceived lack of ease of use on mobile devices.

This study investigated the current method of text selection and interaction for mobile document writing. The study determined that there are shortcomings present in the current method of text selection from both a practical and user experience point of view. Users are much more comfortable with traditional selection methods that utilise familiar inputs such as mouse and keyboard. These input methods were investigated to create a selection analogue for mobile devices. The desktop-inspired features promoted faster and less error-prone text selection, maintaining a positive user experience with a minimal learning curve.

Using the desktop analogue, a new selection method, referred to as “Writing Toolbox”, was designed, developed, and tested, which outperformed the current selection method. The proposed method was tested through practical usage tests and a user experience survey. The results of the tests were promising; showing an average text selection speed improvement of 52.9% while also reducing overall error frequency by more than 59%. The reported user experience improvement showed a 96% preference for the proposed method. Participants of the survey reported an increased incentive to make use of mobile devices for productivity tasks if the proposed selection method was available for use in the software suite. This indicated that the quality of available tools is a significant motivator for mobile productivity software adoption.

Lessons learnt from the development and testing of the current and proposed selection methods are described. Additional shortcomings, identified by the user experience survey, are provided as future areas of study. The provided guidance aims to enhance mobile productivity software and in doing so improve user adoption rates and provide a more intuitive and performant text selection method.

(4)

A new method for content selection on a mobile device ii

TABLE OF CONTENTS

ACKNOWLEDGEMENTS ... i ABSTRACT ... ii TABLE OF CONTENTS ... ii LIST OF FIGURES ... iv

LIST OF TABLES ... vii

1. INTRODUCTION ... 2

1.1. Preamble ... 2

1.2. Background on mobile productivity ... 2

1.3. State of the art (selection methods) ... 10

1.4. Need for the study... 18

1.5. Overview of this dissertation ... 19

2. METHODOLOGY ... 21

2.1. Preamble ... 21

2.2. Investigation of existing methods ... 21

2.3. Development of a new selection method ... 26

2.4. Selection method testing ... 34

2.5. Summary ... 41

3. RESULTS ... 44

3.1. Preamble ... 44

3.2. Selection method test analysis ... 44

3.3. Validation ... 66

3.4. User questionnaire findings ... 70

3.5. Summary ... 80 4. CONCLUSION ... 83 4.1. Summary ... 83 4.2. Discussion ... 84 4.3. Future work ... 89 BIBLIOGRAPHY ... 92

APPENDIX A (PRACTICAL TEST DATA) ... 98

APPENDIX B (USER EXPERIENCE SURVEY) ... 105

1. Blank survey ... 105

2. User comments ... 110

(5)

A new method for content selection on a mobile device iv

LIST OF FIGURES

Figure 1 The blocking effect produced by the keyboard and popup menu. ... 25

Figure 2 Editing and styling interfaces compared. ... 29

Figure 3 The logical steps taken to select a word/phrase/block of text. ... 29

Figure 4 Gesture feedback for a drag selection while in the selection mode. ... 30

Figure 5 The background process of selecting text. ... 31

Figure 6 Normal distribution of single-word selection time (with errors). ... 45

Figure 7 Normal distribution of single-word selection time (without errors). ... 46

Figure 8 Styling menu becomes transparent while the user performs a gesture. ... 48

Figure 9 Normal distribution of single-word selection error frequency. ... 48

Figure 10 Selection time vs error frequency. ... 49

Figure 11 Normal distribution of phrase selection time (with errors). ... 50

Figure 12 Normal distribution of phrase selection time (without errors)... 52

Figure 13 Normal distribution of phrase selection error frequency. ... 53

Figure 14 Selection time vs error frequency. ... 54

Figure 15 Normal distribution of sentence selection time (with errors). ... 55

Figure 16 Normal distribution plots of sentence selection time (without errors). ... 56

Figure 17 Normal distribution of sentence selection error frequency. ... 57

Figure 18 Selection time vs error frequency. ... 58

Figure 19 Normal distribution of paragraph selection time (with errors). ... 59

Figure 20 Normal distribution of paragraph selection time (without errors). ... 60

(6)

A new method for content selection on a mobile device v

Figure 22 Selection time vs error frequency. ... 62

Figure 23 Total errors made (proposed method). ... 63

Figure 24 Total errors made (current method). ... 65

Figure 25 Method ease of use scores from survey. ... 71

Figure 26 Method ease of use scores from survey (single-word). ... 71

Figure 27 Method ease of use scores from survey (multiple word). ... 72

Figure 28 Method ease of use scores from survey (special gestures). ... 73

Figure 29 Frequency of use of adjustment markers vs. ease of use of the markers. ... 74

Figure 30 Preference for menu position (fixed vs. moving). ... 75

Figure 31 Preference for interface separation (combined vs. separate). ... 77

Figure 32 Ease of switching platform (Android to iOS) and continuing to use method. ... 78

Figure 33 Participants’ preference for either the current or proposed selection method. ... 78

Figure 34 Willingness to use mobile productivity tools after (addition of proposed method).79 Figure 35 The blocking effect produced by the keyboard and popup menu. ... 116

Figure 36 Editing and styling interfaces compared. ... 116

Figure 37 Gesture feedback for a drag selection while in the selection mode. ... 117

(7)

A new method for content selection on a mobile device vii

LIST OF TABLES

Table 1 Key areas for software to ensure maximal usefulness. ... 11

Table 2 Computer and mobile selection analogue (current method). ... 22

Table 3 Proposed method design goals ... 26

Table 4 Computer and mobile selection analogue (proposed method). ... 33

Table 5 Practical tests performed, information recorded, and motivation for each test. ... 36

Table 6 User experience survey questions and gathered information explained. ... 40

Table 7 Single-word selection test results summary. ... 45

Table 8 Single-word selection without error results summary. ... 46

Table 9 Phrase selection test results summary. ... 51

Table 10 Phrase selection without error results summary. ... 52

Table 11 Sentence selection test results summary. ... 55

Table 12 Sentence selection without error results summary. ... 56

Table 13 Paragraph selection test results summary... 59

Table 14 Paragraph selection without error results summary. ... 60

Table 15 Selection errors made by users while testing the proposed method. ... 64

Table 16 Selection errors made by users while testing the current method. ... 65

Table 17 T-test results for timing data and average error frequency per selection. ... 69

Table 18 User error data for the current method ... 98

Table 19 User selection timing data for the current method ... 98

Table 20 User error data for the proposed method ... 99

Table 21 User selection timing data for the proposed method ... 99 Table 22 Selection time improvements when switching from current to proposed method 100

(8)

A new method for content selection on a mobile device viii

Table 23 Summary of the group timing data for the current method ... 100

Table 24 Summary of the group timing data for the current method ... 100

Table 25 Selection time improvement summary ... 100

Table 26 T-scores calculated by comparing the summary datasets ... 101

Table 27 Summary of error occurrences during the practical tests ... 101

Table 28 Per-participant selection timing data (including errors) ... 101

Table 29 Per-participant selection timing data (excluding errors) ... 102

Table 30 Per-participant error average error per selection ... 102

(9)

A new method for content selection on a mobile device 1

CHAPTER 1

Introduction

This chapter provides an introduction to mobile text selection. It highlights the shortcomings of the current state of the art selection method in order to derive a set of requirements for an improved replacement method.

(10)

A new method for content selection on a mobile device 2

1.

INTRODUCTION

1.1. Preamble

In the modern world, the need to be productive from any location has become a necessity. This is evident in the rise of many technologies that assist people to continue working even while away from their office. With the rise of the internet and an advanced, connected, and technologically empowered society, productivity is set to keep improving.

More and more people are using off-site productivity tools to keep track of tasks and produce work-related documents. With the increase in the use of these technologies, it is important to understand and produce quality software that improves user experience and ability. User experience refers to the overall experience of a person using a product or software [1]. In the context of this dissertation, user experience describes the ease of use and user satisfaction associated with a given selection method. The user plays a critical role in the development of new software. Utilising user feedback during the planning and testing phases ensures that the software that is produced solves the intended problem. Novel software can also be catered to meet specific user needs by introducing users to the software earlier during the development phase.

This section focuses on the history of mobile productivity and document creation. It aims to show how the average person would use the software and where current solutions [2] are lacking. It also highlights some key areas for investigation that can improve the user experience with document creation on a mobile device. The intention of this simplification is to get more users to employ their mobile devices to complete productivity tasks.

1.2. Background on mobile productivity 1.2.1. Offsite work and education

Productivity has been a significant issue for as long as work has been done [3]. In his paper, “The research of service enterprise productivity and its influencing factors,” Li discusses four proposed factors that describe productivity in the service enterprise environment. These factors are strategy, customer, operation, and technology. As these factors focus on the service environment, the only truly applicable factor for this study is technology. The technology factor is described as the way the development and application of internet and computer technology influences the worker; or in this case, the user.

The use of appropriate technology in industry has already been shown to greatly improve productivity [4]. This is clear in industrial examples where manufacturing and resource

(11)

A new method for content selection on a mobile device 3 production is concerned. Entire factories are composed of robots assembling cars, canned foods, and other products with little to no human intervention. The challenge, however, is that this same automation is difficult to achieve in the fields of more cognitive and creative work. Therefore, a shift in focus is needed to create tools that allow workers in these fields to improve their own productivity by completing tasks faster and more efficiently while maintaining the quality of their work.

In recent years there has been a large focus on improving the productivity of off-site or field workers using mobile technology [5]. This improvement is not simply based on the tools that the field worker will need, but also the support that the field worker should receive. To this end, models have been produced that attempt to describe mobile workers and their needs, how their needs differ from that of office workers, as well as how to better design and develop tools and support for them [6].

The need for these tools is clear when considering the situation that the field workers find themselves in. Some work can simply not be done off-site, which means that potential productive time is lost during the commute to and from site. Using the tools that mobile productivity provides, in many cases, it would be possible to regain the lost time by completing small tasks during the commute rather than only starting to work when one gets to the office. The idea of being able to continue to work during previously unproductive times is what has renewed the interest in mobile productivity technology.

A key area where mobile productivity has seen broad adoption is in schools and universities [7] [8]. Members of generations after the Millennials use smart devices on a regular basis, are familiar with the device capabilities, and are now also using these devices for work rather than simply social interaction and games [9]. Technologies frequently used by students are that of time scheduling and document creation applications. These applications let students use their smart devices to create and edit their work in absence of a computer [10].

Some of the most commonly adopted applications by students include the Google “G Suite”

1

of applications and their desktop counterparts, the Blackboard2 communications application and Trello3 for time management. The “G Suite” set of applications allow the creation of standard text-based documents (Docs), slideshows that can be used for presentations

1 Found at: G. Cloud, “G Suite,” Google, 2019. [Online]. Available: https://gsuite.google.com/intl/en/.

[Accessed 07 03 2019].

2 Found at: B. Inc., “Blackboard Learn,” Blackboard Inc., [Online]. Available: https://www.blackboard.com/blackboard-learn/index.html. [Accessed 07 03 2019].

3 Found at: Atlassian, “Trello,” Atlassian Pty Ltd, 2019. [Online]. Available: https://trello.com/home. [Accessed 07 03 2019]

(12)

A new method for content selection on a mobile device 4 (Slides), spreadsheets that can be used for financial or data analysis tasks (Sheets), and allows the creation of dynamic forms with which the user can gather data (Forms).

Blackboard is a collaboration and education tool that allows instructors to upload and share schedules, materials, and tests with the students in a user-friendly web platform and mobile application. Trello is a task management application that allows workflow and tasks to be monitored in order to improve productivity and awareness when working alone or in a group setting. All these applications have the theme of simplicity, with one common goal: giving the user the information and tools needed to complete the task at hand without overwhelming the user with features.

The use of mobile applications in the classroom has the added benefit of allowing students to engage and collaborate more than without these technologies [11]. The students also showed an improvement in attitude towards work and learning when exposed to the technology-assisted environment [12]. To take part in this new method of education and collaboration, the user must first possess an appropriate mobile device. The selection of which device the student should use is based on various factors. The most important of these factors include size, software diversity (what applications are available), and ease of use, indicating that the users want a simple and fluid experience with a minimal learning curve [13].

1.2.2. User experience and testing

Studies have shown that users tend to prioritise the ease of use of software over the functionality thereof [14]. This indicates that at times, useful software could be cast aside due to the difficulty of use or a complex user interaction method. The ease of use of the software can be defined as how easy or convenient it is for the intended user to interact with and use a system [15]. In the context of this dissertation, the system refers to the selection methods (proposed and currently available).

The ease of interaction of the software is influenced by multiple factors, such as the layout of the software user interface, input methods available to the user, feedback mechanisms (visual and mechanical), and shortcut actions (macros and prediction) [16]. The interactions mentioned are only some of the aspects that the software designer can look at to improve the user experience, and subsequently the ease of use of the software.

The task of designing an easy-to-use, responsive, and appealing user interface for an application can be daunting. The metrics with which the developer can measure the usefulness of the user interface are not always tangible or empirically measurable; therefore, user interaction testing is needed with real users.

(13)

A new method for content selection on a mobile device 5 Users can test the application in ways that were perhaps unexpected to the developer, which would not have otherwise been found as the developer knows how to use the application and how to avoid errors. This type of testing can be augmented with automation tools that allow a certain level of testing before the users are exposed to the application [17].

Using the automation tools and guidelines for good user interface design, an ease of use rating can be determined for the software with a reasonable amount of accuracy. The Google “Play Store” employs a similar system when it comes to the testing of applications that have been submitted to the store. An automated system launches the application and attempts to interact with the application by identifying and interacting with different elements on the page. This is analogous to a new user trying everything in the application when they use it for the first time. This system tests a broad scope of interaction, but the majority of the testing is done on the touch interface and how the application performs with different gestures.

1.2.3. Mobile productivity 1.2.3.1. Document writing

When concerned with mobile productivity applications, the most widely used applications fall into the category of writing and communication. This includes applications such as Google Docs, Word Mobile, and WPS Office. These applications focus on the compilation of documents on a smart device which can then be transferred to a computer. They are built to allow many of the same features of the desktop version on the mobile version.

The task of prolonged text input associated with productivity/document creation on mobile devices is not optimal in all cases [18] [19]. Many mobile productivity applications emphasise the writing aspect of document creation, which is the core of what a user must do to create a document. A large amount of research has been done on the viability of people using mobile input methods to write long pieces of text. It has been found that pure text entry is slower than that of writing on a desktop [20] [21]. The cited causes for this include the smaller keyboard, the lack of tactile feedback, the inability to use all one’s fingers when typing, and the screen size that forces the screen to move to keep focus with the text [16] [21] [22].

1.2.3.2. Text prediction for writing

Some of the current solutions to the problem of text entry can be found on most modern smart devices. The first and most common method of handling the problem of text entry on mobile devices is that of predictive text [23]. This method, also known as auto-complete, involves the analysis of the text that the user is typing to make suggestions as to the most

(14)

A new method for content selection on a mobile device 6 probable next word that the user will be typing [24]. This is commonly encountered when typing on a smartphone but is also present online when using search bars, and is suggested for use on desktops to assist with writing for slow or otherwise impaired users [25].

Studies have shown that users who use predictive text can save a lot of time by employing this strategy when composing text on a smartphone, with users who type more slowly showing the highest gain in speed [26]. Predictive text has the added benefit that a user can be prompted with the correct grammar and spelling and thus elevate the standard of language with which they communicate.

Predictive text is an effective way to increase typing speed, mitigate errors, and improve the normal input method; however, most predictive text implementations rely heavily on previously typed text [24]. The predictive text models that are present on the majority of smartphones seldom predict words that the user uses infrequently. These prediction systems are less useful when the user engages in a new topic of conversation or writing. Predictive text also regularly suggests simpler language than what the user may have intended to type but opted for since it was suggested. This has been shown in some studies to create a sublanguage that is mostly used on mobile devices that is more concise but less formal, a non-ideal situation if the user is writing work-related documents [27].

1.2.3.3. Dictation and Speech-to-text

Voice recognition technology has also advanced in the last few years and has become a viable method of input on mobile devices [28]. This method of input is referred to as dictation and is most notably found in the various mobile assistants present on modern phones. These assistants include Siri (iOS), Cortana (Windows), Google Assistant (Android), and Alexa (Amazon). The use of these assistants to complete simple tasks on the mobile devices has shown their potential as an input method which can alleviate the spatial issues connected with the on-screen keyboards present on mobile phones.

Dictation is still not a perfect solution [29]. The accuracy of voice recognition systems can degrade rapidly depending on the environment in which these systems are used [30]. Voice dictation is sensitive to external noise and can produce various recognition errors if the ambient noise becomes too loud. This can be mitigated by only using this system in relatively quiet environments. This means that its utility is lost in loud environments or public spaces.

Furthermore, dictation is a burden in that most of the assistants require a working network connection to function as the voice processing is not done locally on the mobile phone itself

(15)

A new method for content selection on a mobile device 7 [31]. Using the system will thus cost the user money in the form of data, which the standard typing interface does not. For use in creative works, dictation and the tempo of speech have their own drawbacks.

The services are executed server-side and thus the raw audio data needs to be uploaded for processing. The service also controls the duration of the dictated recording and when to stop listening for input [32]. This is problematic as the user tends to pause to better structure their idea when performing creative or cognitive writing (dictation) tasks. This could trigger the assistant to stop listening for input during the pause to process the audio, losing continuity in input and frustrating the user.

The final hurdle that voice recognition must overcome in order to be usable as a robust input method is that of accents. Many nations speak English, but some have drastically different ways in which certain words are pronounced. This variation in speech makes the detection of words difficult depending on what data set the assistant was trained in to detect the given language. The possible solution for this is to locally retrain the assistant to recognise the user’s voice with a small training set of phrases that the user can say to modify the assistant for better recognition of spoken words [33].

The benefit of detection being executed server-side is that significantly more powerful machines can perform the detection, with better resources and larger data sets. This also allows for the possibility of detection to be tailored to the specific user if a profile is stored, increasing accuracy for the user who frequently utilises these systems [34].

1.2.4. Touch screens and user interaction 1.2.4.1. Screen size and keyboard input

Screen size is a common factor that users identify when questioned about their use of mobile devices for productivity. Some opinions that arise are those of the screen being too small to be typed on for extended periods of time, too small to get overall context in a document, and that the screen resizes frequently which could inhibit productive workflow. This problem could be further compounded insomuch that with modern touch screen devices, a portion of the screen is used for the input method. This reduces the screen size by the height of the software keyboard, and in some of the worst cases (when using a small screen or horizontal orientation), could block almost the entire screen, making accurate, sustained input almost impossible.

When users are busy with productive writing in an academic environment, the use of external sources is frequently part of the writing process. The reliance on research and

(16)

A new method for content selection on a mobile device 8 retyping of the findings in other sources also shows up as a problem when the small screen of a phone is considered. The reduced size of the phone screen has the effect that in most cases a second document cannot be shown alongside the one the user is currently editing. The lack of direct access to the source material while typing slows down the typing process as the user must frequently reference the material to accurately use it in their writing. This problem is made worse when the information regards a technical topic.

There are ways that the problem of a small screen space has been countered. The use of dynamic (changing based on input type) and dismissible keyboards is present on many smart devices [35]. These keyboards allow less of the screen space to be occupied by the keyboard, or to completely dismiss the keyboard if not in use [36]. This allows a larger viewport to the actual document. The user can also switch between the different documents (applications) with relative ease using the application dock; however, this comes with the problems as mentioned at the start of Section 1.2.4.1.

Users with larger screens, such as those that come on modern flagship phones and tablets, can also make use of the dual application modes in which two applications can be displayed at the same time. However, this again encounters the problem with screen space, where the second application, as well as the keyboard, now limit the space that the primary application can occupy on the screen.

1.2.4.2. Touch gestures and pointing

When considering productivity on mobile devices many people overlook one of the most basic input tools that is missing from the standard productivity suite available on a desktop computer, namely the mouse. The lack of a mouse has the effect that much of the functionality must be layered behind various trigger mechanisms and handled in a more complex way since the dedicated selection and pointing tool has been removed.

The pointing functionality of the computer mouse has been replaced on smart devices using the touch input in the form of a tap, which actuates buttons, sets cursor focus, and acts in the same manner as a mouse left-click. A tap on the screen can act as both double-left-click and double-right-click, either selecting some short text or showing a context menu, respectively.

In general, the dragging of the user's finger over the screen acts as the navigation method of the scroll wheel on a computer mouse, with regular finger movement acting as mouse movement of the cursor. With this description, it is obvious that the interaction with the smart device’s screens are heavily analogous of that with which users are used to interacting with on a computer screen, with small variances due to the nature of the device they are using.

(17)

A new method for content selection on a mobile device 9 The problem arises when the user attempts to perform more complex actions other than simple navigation and setting focus on a text document. The actions of selecting text, triggering context menus, trying to cut, copy, or paste all come with a time delay while the user intent is determined [37]. The user intent determination step is necessary due to the integrated nature of the text editing and text manipulation interfaces, which form two parts of the document creation view.

1.2.4.3. User intent considerations

Integrated text editing methods need to determine user intent. This is due to an overlap in the selection and navigation actions’ gestures. Time is lost while the software determines user intent and errors can arise from incorrect intent recognition.

A common example of the necessary delay to detect user intent is that of a user trying to select some text using a press action. When trying to select some text using a press action, the system initially detects a screen touch which could potentially trigger a tap event. The system has to wait for a touch-end event from the screen to fire before it can determine that a tap has occurred, or it has to wait for a given interval (300ms, in the case of native gestures for Android) to determine that the gesture was not a tap but a press gesture which can now select the word. This wait time where the system is effectively idle can hinder the user since the immediate intent is not known before the full gesture is complete.

This is also shown in the common way users select short phrases in a text area. Unlike the desktop method where the left mouse button can be held and dragged to select text, the user must make a single word selection first. The selection can then be adjusted to select the length of the entire phrase using software “teardrops” that appear upon selection of the initial section of text. The problem of time delay is encountered again, with the compounding issue of multiple gestures needed to perform a single task, introducing potential error sources.

With releases of Android 9 (Pie) and iOS 9 operating systems, it has become possible to select text in a more direct manner. Android allows the user to double-tap and then swipe to select text, but the inherent delay is still present, as can be noted with the selection feedback taking the same amount of time to appear as with a press action.

iOS implements a different solution, with the inclusion of force touch to immediately start a selection of text (supported devices being the iPhone 6S to the XS Max). This is a possible solution which could work well given that the device the user is performing the selection on supports the force touch technology. It should be noted that with the release of iOS 13, force touch (also known as 3D touch) has been replaced with haptic touch on the latest model

(18)

A new method for content selection on a mobile device 10 Apple devices. Force touch thus does not provide a solution that can be made universally available to all devices to solve the problem of easy and fast text selection.

1.3. State of the art (selection methods) 1.3.1. Current input methods

The current selection methods that are available on smart devices vary between the operating system and device hardware. Some selection methods rely on auxiliary input devices (stylus [38]) to be used in conjunction with the smart devices, while most make use of the standard set of features (multi-touch [39] or hardware buttons) of most devices and operating systems.

Common input methods that currently exist on smart devices are that of touch interactions and gestures. These interactions and gestures have been used and standardised across devices to allow users to select, edit, and manipulate text using the touch interfaces present on modern smart devices.

The other input methods using third-party hardware usually present themselves in the form of external keyboards and gesture input devices such as digital styluses [39] which can also be used for writing on smart devices. The use of external keyboards brings with it the use of an arrow key for navigation and inline text selection that is not common on mobile smart devices.

Even with these varied input methods and devices, extensive research has been done to present novel input methods using the hardware available [2] [22] [40]. The goal of this dissertation was to find more adept solutions to the problem of smart device input, and in the case of this dissertation, text selection. In this section, these various methods will be explored to illustrate the varied nature of input methods, how these methods can be applied to the text selection problem, and to determine the key areas to measure the effectiveness of an input method for text selection applications.

1.3.2. Text selection requirements

One of the first aspects to look at when concerned with text selection is the fundamental needs of text selection where it will be used. Writing in any language can be viewed as a creative activity, which would necessitate the editing, review, and rewriting of sections of the compiled piece of text [41]. Using the application of text selection in the writing environment as a guideline, appropriate methods can be developed to create tools that will aid with selecting text and manipulating the selected content.

(19)

A new method for content selection on a mobile device 11 The guidelines for how writing is done on devices stem from research as early as 1977, with Chamberlin positing the intricacies of text editing in his paper "Text Editing". He notes the differences in editing for natural language text writing, text editing for programming purpose, and touches on other concepts such as scrolling, indexing, and word wrapping [41].

Chamberlin also notes that creating editing tools for natural text is more difficult, as the code was mostly (at the time of his writing) line-based. The line-based nature of code meant that lines could be treated as distinct units, where lines have little meaning when looking at natural text. In natural text sentences, paragraphs form the distinct units, but Chamberlin states that a good English (natural language) editor should be able to handle programming as well since programming editing is less complex.

Taking Chamberlin's concepts into account it becomes obvious that selection methods should be limited to logical text sections. This means that text selection is focused on selecting logical text blocks such as words, phrases, sentences, and paragraphs. It is unlikely that the user would want to select single characters within a word. Sentences and paragraphs can easily be distinguished using punctuation and spacing as guides, with phrases being the most diverse selection that should be catered for.

Sharma and Gruchacz further build on the concepts of text editors and how the ideal editors should be implemented [42]. In their paper, “The Display Text Editor TED: A Case Study in the Design and Implementation of Display-Oriented Interactive Human Interfaces,” the aim was to determine criteria with which one could design and improve human interfaces in software. The discussion in this paper is useful as it focusses on text interfaces with extra input from pointing devices, a precursor to the problem addressed in this dissertation.

In the study, five key areas are identified in which software should excel in order to be maximally useful for the end-user. The areas determined were ease of learning, remembering and use, safety, and customisability, as noted in Table 1.

Table 1 Key areas for software to ensure maximal usefulness.

Number Area name Discussed

1 Ease of use Section 1.3.3

2 Safety Section 1.3.4

3 Customisability

4 Ease of learning (adoption) Section 1.3.5 5 Ease of remembering (familiarity)

(20)

A new method for content selection on a mobile device 12 The areas and their context in the editor software that was developed are discussed, as well as how the software was improved by incorporating these aspects.

1.3.3. Interface simplicity

Building software based on these key areas, with text selection as the goal, certain design philosophies that should be followed become obvious. Firstly, the concept of keeping the interfaces as simple as possible. This does not mean that only the bare minimum features are present, but rather that only the features that the user needs should be present.

This can be seen in the writing and editing context with a standard style menu. Bold, italics, and underline are top-level actions that most users will use at some point and thus are positioned to be easily accessible. These features can also only be shown once a selection of text that can be styled is made. Hiding options until needed keeps the user interface clean and avoids menus obstructing user actions.

Keeping the interface simple also means that the user can easily learn the environment and come back to it with minimal effort. As stated in Sharma and Gruchacz’s paper [42], there is a fine line between adding power for the user and adding complexity. The user should at no point feel that they are overwhelmed with the actions available to them.

1.3.4. Safety and customisability

Other benefits of simple interfaces incorporate the safety and customisability aspects. If the user is limited to what they can do per interface (or mode, as it is called in the paper) then the user’s ability to make mistakes is also reduced. The effect of this is that mistake occurrences and severity are reduced, with handling being implemented to help the user remedy their mistakes. In the context of text selection, this could mean that when selecting text, the user can only style or copy the text, not type it. This could avoid accidental deletions, overtyping or moving of text sections (blocks), and provides handling to reset the styling of selected text if mistakes were made.

Some of the major functions of selecting text are to use the linked functionality of cut, copy, and delete. The use of these features has become second nature to those working in text-heavy environments. The ability to quickly copy a paragraph or phrase, cut a section to move it or delete it completely without the need to retype or use backspace improves workflow by minimizing retyping the same text or waiting for the cursor to run back and fully delete the given page.

There are, however, complexities that must be considered with cut, copy, and delete. Mann discusses the cut, copy, and paste operations (delete excluded) in which he notes the

(21)

A new method for content selection on a mobile device 13 possibility of inconsistency that arises with the use of these operations [43]. The paper suggests a more complex set of commands that can be used to avoid inconsistency and keep the functionality that the user expects.

1.3.5. Familiarity and learning curve

These suggestions move away from the notion of keeping adoption of the systems familiar and easy to learn, aspects regarded as important to reduce the learning curve of software [42]. Taking both ideas into account it can be said that the action that the user wishes to perform should be rooted in what the user expects from current methods but still allow the user to specify the fuller action that they wish to perform.

In a modern gesture-based interface these varied actions can also be linked to custom gestures that the user can perform. This allows the user to select the text as needed, and with minimal further effort, perform the required action without the need to interact with a menu. This keeps the interface simple, builds on the user’s current actions of selecting to continue and not switch interaction methods, and allows actions with nuance without the need for complex decision menus.

With the knowledge gained from the previous studies on what users need to be able to select and edit text, a new method can be devised to improve upon the current method. Multiple studies have attempted to create new input interfaces for mobile devices. These new input methods could improve on selection time and accuracy if implemented, and provide alternatives to the currently adopted input methods. The rest of this section focuses on the review of these methods to determine if they have application in mobile text selection.

1.3.6. Previous methods 1.3.6.1. MoCamMouse

This first study to consider here is that of the “MoCamMouse (Mobile Camera-based Mouse-like Interaction)” [40]. The study aimed to determine whether using a camera-based input for mobile devices could solve some of the traditional problems with mobile touch interaction. The two largest problems the authors listed that the study tried to address were that of hand occlusion and physical contact with the device that could lower the user experience.

For the study, the “MoCamMouse” software was developed and tested. This method uses the camera that is present on the mobile device to perform gesture recognition of the user’s hand on a surface. The application interface is overlaid with a ghost image of the user’s hand which allows that user to interact with the application without their hand obstructing the interface. The user can perform traditional mouse-based actions such as left-click and

(22)

right-A new method for content selection on a mobile device 14 click by using the remapped volume buttons on the side of most modern smart devices. This allows the user to left-click via a volume up action and right-click via a volume down action. Long press could also be achieved pressing the volume up action for a short period of time. The solution was used in application to perform actions using the finger as a pointing device (while wearing a fingertip tracker tip), as well as outside of specific applications, and could be used to interact with the operating system and to launch applications. Thus this solution is generic and more useful since it is not application-specific. The authors determined via questionnaires that the users found the method to be intuitive and easy to use while minimising the problems of occlusion and physical contact with the screen.

This method is interesting as it proposes a different use of some of the hardware that is currently available on the mobile device. The method could find good application in text selection as hand occlusion is a large problem for many users, specifically those with larger fingers.

The drawbacks of the method, however, make it unviable for use in an editing environment. The user must pick up the device to enable the use of the camera, which, in a scenario where multiple selections must be made, could lead to the device frequently transitioning positions. Furthermore, the holding of the device to use this method could lead to fatigue in the user which means that they would be less prone to use this method over current methods.

1.3.6.2. Virtual trackball

Other interaction methods that also work based on the back-of-device input is that of a virtual trackball. The virtual trackball could be used as a pointing and scrolling device, leaving other touch interaction on the touch-enabled screen to be used for other functions. Applications already exist that can extend the functionality of the fingerprint sensor of modern smart devices to accept various gesture inputs.

“Fingerprint Gestures” is an application that relies on Samsung and Google fingerprint APIs and enables single-tap, double-tap and swipe gestures using the rear fingerprint sensor [44]. Another implementation of this type of interaction is discussed by Kratz and Rohs in their 2010 paper, “Extending the Virtual Trackball Metaphor to Rear Touch Input” [45]. In the paper, Kratz and Rohs attempted to determine whether a fully realised virtual trackball would be achievable and be more usable with manipulation of software and models in a 3D context.

(23)

A new method for content selection on a mobile device 15 Kratz and Rohs proposed achieving the input method using what they referred to as an “iPhone Sandwich”. This “sandwich” entailed two iPhone smartphones attached back-to-back to allow input on the front and back-to-back of the complete device, thus allowing a full virtual trackball to be achieved. Allowing back and front interaction and using the full-featured screens meant that a more gesture-complete input method could be achieved, allowing more varied input combinations. Kratz and Rohs also conducted a user study in which they determined that the full virtual trackball worked better in the manipulation of objects than the current solution of tilting the devices to achieve manipulation.

For text selection, the same benefits are seen such as that with the “MoCamMouse” [40] implementation. The rear input would allow the user to interact with the screen content without the hand occlusion by using the rear panel. The user could use the fact that the rear panel could accept various tap interactions to perform more complex gesture combinations, allowing drag-selections without time delay, for example.

Some menu interaction could also be added to the rear interaction, allowing popup and context menus to be accessed via taps on the rear screen, permitting a simplified and less cluttered user interface on the front panel. The same problems as with the “MoCamMouse” exist here as well, where the device must be held in the air to reach the backside and make use of the functionality. The user might also be forced to use a non-standard grip on their device which could mean less control and an increased risk that the device may be dropped, which makes this method of interaction undesirable.

Input methods have now been broadly discussed outlining the pros and cons of each method and how these methods could be applied to text selection. The following sources have applied a variant or similar method to those discussed above to a text selection problem. The results of these studies are discussed to try to determine which text selection problems these methods solved and whether a combination of these methods could provide a complete, multi-platform solution.

1.3.6.3. ‘Pen+touch’ gestures

In their 2013 study, “The Study of A Novel Gesture Technique Based on ‘Pen+touch’ in Multi-touch Interfaces” [22], Cheng and Yin describe a gesture method where a stylus is used in conjunction with finger-touch input to perform actions on a smart device. The goal of the study was again to alleviate what the researchers called the “fat finger” problem by using the precession provided by the stylus along with the possibility of extra control over gestures using the hybrid touch approach.

(24)

A new method for content selection on a mobile device 16 This approach allowed the expansion of the possible number of gestures that could be recognised. The team had various input combinations in which basic input, gestures, and unistroke commands of the fingers and the stylus were recognised. Using the different input methods, thousands of input combinations could be achieved, which could mean the software interface could be free of unnecessary buttons in favour of combination gestures. The study conducted a user experiment in which users were provided with tasks to complete using the gestures and detailed instructions on how to do so. The results that were obtained indicate that even though a plethora of gestures were available, the simpler gestures were still the better choice. The simpler the gesture the less likely the users were to make mistakes, as well as the fewer fingers used in the gesture combination. The results also indicated that gesture combinations that did not involve stroke recognition were more ideal and easier to use.

The findings that simpler gestures perform better and are more user-friendly are supported by a user study done by Schenk as part of his paper “Approach to Gesture-based Editing of Diagrams” [46]. Schenk also notes that gesture usage is more akin to the traditional pen and paper environment and is best utilised in support of this method of content creation.

Cheng and Yin’s study shows that the gesture methods could find application in text selection to act as a gesture modifier. Where currently multiple-finger input has little use when selection and manipulating text is concerned, the use of combination gestures allows special actions such as that of a “shift” key. This alternate mode could change the behaviour of standard gestures, allowing reuse of these gestures for other tasks. An example would mean that while no touch is detected a drag action scrolls the page, but while a touch is detected the drag action becomes a selection tool.

1.3.6.4. Keyboard localised gestures

A 2013 paper titled “Gestures and widgets: performance in text editing on multi-touch capable mobile devices” [47] also gave a new way of selecting text that builds on then-current (2013) selection methods. The paper aimed to design and implement a new gestural method of text selection. The developed method could be added as a service and thus be used with any text editing task on mobile devices running the Android operating system. The method was based on gestures that were done on the software keyboard to select and manipulate text. The cursor would be placed in the text body and gestures, such as the two-finger swipe to select text backwards, would be incrementally (word-based) performed. After a selection was made, further gestures could be used to cut, copy, or paste the text.

(25)

A new method for content selection on a mobile device 17 The ability to use gestures to perform these actions meant that there was no need for the context menu, which frees up user interface clutter and improves timing. The study found a 13-24% performance benefit when using the gestural method to perform the selection and editing actions. The gestures also provided an important action that is not always available on mobile devices, which is the ability to move the caret (typing focus indicator) in character steps similar to using arrow keys. Moving the caret is a great help in text editing as it allows focus-shifting in the text body without hand occlusion, which has been noted in Section 1.3.6.1 and Section 1.3.6.2 as a problem in editor interfaces.

Using these gesture methods could provide benefits such as improving ease of use, editing accuracy and speed, and minimising context menu use in the UI. The true power of the method described is that it can run alongside the default methods. The user thus has multiple options to complete a task and can choose which methods they prefer, using only the gestures that they are comfortable or see improvement with.

1.3.6.5. Styluses and enhanced gestures

Editing tasks on modern devices can also be augmented using hardware that is available with these devices. One piece of hardware that eases text selection is a stylus. A stylus is generally slimmer than the finger of the user, thus limiting the occlusion effect. The tip of the stylus is also small and allows more precision than achievable with a finger. Stylus use also benefits from the training most people receive when learning to write, as the use of a stylus is extremely similar to the use of a pencil. Stylus-based editing presents an effective way a user can select text with precision.

Costagliola et al. use this method of editing in their study to expand upon gesture editing, using the enhanced precision to add to the list of gestures usable to modify the text body [48]. The study employed a user study to first identify the ideal gestures for text editing and manipulation by users. The results of the user study were employed to design a gesture method to edit a natural body of text. This method was evaluated by a further user study and found to be favoured by the users and to be more performant than current methods.

The methods of this study could be positively applied to the development of fully realised text selection and modification interfaces. The ability to perform most editing tasks using gestures empowers the user and increases editing speed and accuracy.

Gesture selection using a stylus can have unexpected benefits, as shown by Doreswamy and Mittal in “Improvement of Power and Time by gesture-based Text Selection in Mobile Web Browser” [49]. Their study attempted to design a method of selection that would improve upon the time taken and power consumed during text selection actions.

(26)

A new method for content selection on a mobile device 18 Doreswamy and Mittal developed a selection method that relied on the use of a stylus that came equipped with a hardware button. The selection process consisted of using a click of the hardware button to, and tapping the stylus on, the desired location to start the selection. The gesture is repeated to indicate the end of the selection. The user can scroll the text body during the selection process. After this process, the text between the start and end points would be selected and could be edited as if they had been selected normally.

The results of the study showed a mean improvement of 29.06%, 61.07%, and 74.92%, respectively for small, medium, and large bodies of text. The improvement in power consumption found is not relevant to this study. Both above-mentioned studies have the same problem, which is the need for a stylus to be present for the methods to effectively work.

The solution proposed by Costagliola et al. will work without a stylus; however, less effectively, as it requires accurate pointer placement for many of the gestures. The solution developed by Doreswamy and Mittal will not work without a large adjustment to compensate for the lack of a hardware button. Volume buttons might work for this, as seen in the “MoCamMouse” study [40].

From the above studies, there are various lessons that can be learned. A selection method must be easy to learn, once learnt, easy to use, and should function on multiple platforms. Taking the results of the studies and how users reacted to the implementations of the text selection concepts, a specification can be determined for a new method which caters to most of the user needs.

1.4. Need for the study

The literature reviewed, in Section 1.2 and Section 1.3, shows that there have been many attempts to describe and solve the problem of text selection in the past. Studies vary from describing what is needed for text editors and text selection, to studies that attempt to use advances in technology and hardware to improve selection methods. Using the studies as a basis to work from, a new method of text selection for mobile devices with touch interfaces could be devised. This new method should fulfil the following criteria:

1. User-friendly with a shallow learning curve. 2. Faster to select than current implementations. 3. Multi-platform compatible.

4. Simple error correction (avoid teardrops if possible). 5. Hardware independent.

(27)

A new method for content selection on a mobile device 19 Current solutions do provide some of the above-mentioned characteristics, but in most cases, every solution falls short in one or more of the aspects mentioned. To be a viable successor to the current selection methods, the proposed implementation should provide a better text-selection experience to the user, not simply an empirically faster method. Better user experience is likely to promote mobile device adoption and use for writing tasks which would normally be performed on a computer/laptop.

The goal of this study was to produce a selection method that fulfils all the above criteria as closely as possible, and compare this method to the current state of the art to determine efficacy and user experience.

1.5. Overview of this dissertation

This section gives a brief overview of all the chapters of the dissertation. The chapter goal is described followed by a short summary of its content to guide the reader to the desired content.

Chapter 1 provides background on the state of mobile productivity in modern society. The

use of productivity tools for writing is discussed. The current state of the art for text selection is introduced to the reader, highlighting the shortcomings of the current method. The shortcomings identified are subsequently used to derive the need for the study of designing an improved text selection method for mobile devices.

Chapter 2 explains the method followed during the development of the proposed solution.

The identification of requirements for a text selection method and the use of current desktop methods to provide a text selection analogue are explained. The method implementation, design decisions, and areas of interest for testing are explained, followed by the testing methodology for both practical and survey-based testing.

Chapter 3 discusses the results gathered from the practical tests. The results were shown to

be positive, with the proposed selection method outperforming the current selection method with which it was compared. This chapter also discusses the practical results validation and user experience survey that was conducted as part of the study.

Chapter 4 gives an overview of the achievements of the study, and discusses the results

and lessons learnt from the practical testing and user surveys. The knowledge gained from the study is summarised and suggestions for future development of selection methods is given to further improve the selection method performance and overall user experience. All relevant test data and results from the conducted surveys are attached as appendices.

(28)

A new method for content selection on a mobile device 20

CHAPTER 2

Methodology

This chapter explains the method followed during the development of the proposed selection method. The design decisions and proposed testing methodology are provided.

(29)

A new method for content selection on a mobile device 21

2.

METHODOLOGY

2.1. Preamble

This section explains the way in which the problem described in Chapter 1 was approached and how the proposed solution was developed. The section describes the existing selection methods available on desktop that were taken into consideration during the development of the proposed solution. The similarities between these methods and current mobile selection methods are explained and transformed into a mobile selection analogue to show how common selection features are implemented.

The shortcomings of the current methods are identified and discussed, using these findings to create a design specification for the proposed method developed during the study. The proposed method (also referred to as “Writing Toolbox”) is described fully in terms of its function and how this method differs from the current methods used on mobile devices. Screenshots are provided throughout this section as visual aid to the selection method, with iOS screenshot variants provided in Appendix C.

Lastly, this section describes the design of the testing procedure, both practical testing and user survey-based. It explains the reasoning behind each test, what the expected outcome would be, and why the results of the given test would be important to capture and analyse. Validation of the method and results gathered are explained in Chapter 3.

2.2. Investigation of existing methods 2.2.1. Desktop computers

When considering the problem of selecting text in a paragraph, the first step was to look at existing solutions. The solutions that were already in mainstream use were investigated to see if there were key concepts that could be adapted to design an improved solution. The most prevalent selection tool in use for document writing is the computer mouse. Pointing devices, such as mouse devices, keyboards and trackpads, form a core part of content selection.

The standard computer mouse consists of a left, right, and middle mouse button (LMB, RMB and MMB), as well as a bottom sensor to detect mouse movement. The MMB dually functions as a button and as a scroll wheel with which to navigate windows too large for the given screen to fully display. These hardware interfaces are all used in the context of text selection and have an analogue available on most mobile devices.

(30)

A new method for content selection on a mobile device 22 The actions performed are expected as the computer mouse has become a standardised input device for computing. Trackpads form a hybrid input device, where some mouse actions, such as left and right-click, can be performed using the hardware buttons or the pad itself. The trackpads thus try to give a ‘best of both worlds’ implementation of both touch implementations and a hardware mouse.

2.2.2. Desktop-mobile analogue

Touch screen devices do away with hardware buttons completely and only rely on touch actions performed on the device’s screen. This is further augmented with various gestures that the user can perform to replace some of the actions that were lost when the hardware mouse was removed. Table 2 illustrates the different actions that a user can perform by means of the computer mouse, and how these actions are replicated on mobile devices.

Table 2 Computer and mobile selection analogue (current method).4

Desktop

interface Action Result

Current mobile analogue

Left mouse button (LMB)

Click Moves the cursor to a new location within the text body.

Single tap (on-screen). Click and

drag

Selects the text over from the start location of the drag to the end location where the button is released.

Various (most common is “Long press” after which the selection markers are moved, by dragging, to desired locations). Double click Selects the nearest full word to

the click location.

Double-tap. Triple-click Selects the entire paragraph in

which the action was performed.

Triple tap, frequently unimplemented. Right mouse

button (LMB)

Click Opens the context menu with various options for the given element or selection.

Press and hold.

Middle mouse button (MMB)

Click Changes the cursor to a scroll allowing scrolling with mouse movement up and down.

N/A

Scroll Scrolls the window in up or down (sideways scrolling achieved by holding the ‘shift’ key).

Finger drag across screen or swipes.

Movement sensor

Mouse movement

Moves the mouse cursor on the screen, allowing other actions to take place at various locations in the text body.

Movement of on-screen elements. In cases where this happens, drag and swipe are used.

(31)

A new method for content selection on a mobile device 23 The analogue discussed above is drawn from the use of the mouse as a pointing device and the way the various actions are implemented on mobile devices at the time of writing. There are other gestures that are used on mobile devices that provide extra functionality that mimics mouse actions.

Pinch and Zoom are examples of these gestures; in this case, being used to zoom in and out of a document. This can be achieved using a mouse scroll wheel when in a zoom mode of an editor or if accompanied by keyboard shortcuts.

The use of the keyboard to further augment the pointing device is a key aspect of its use. Keyboard augmentation allows different selection methods to be used which would otherwise not be available to the user. Selection of a phrase using the ‘Shift’ key alongside single left-clicks at the start and finish locations of the phrase serves as a prime example. Multiple word/phrase selection also becomes possible when keyboard augmentation is used. In their 2013 paper, Cheng and Yin show that it is possible to access a large multitude of gestures when using both hands for input [22]. Their paper shows the power of using multiple hands and, by extension, multiple input devices to perform gesture-based actions. A final method that can be considered is that of using only the keyboard as an input device. The arrow keys along with the ‘Shift’ key produces a ‘click and drag’ analogue that can select phrases. Using this method for selection is common in areas where the use of the computer mouse is less frequent, such as coding.

2.2.3. Error correction

It is clear then that primary selection (first attempt) is done in various ways while utilising different types of hardware to augment the selection methods. The other side of text selection that needed to be investigated was that of selection correction. It is unreasonable to expect any user to always make the correct selection the first time, every time. Considering this, there must be a way to reset or modify the current selection to achieve the desired selection.

On computers, the method that is most used is simply allowing the user to reselect the desired text. This utilises the mouse most of the time, with the user performing the same selection action and trying to select the desired word or phrase on their consecutive attempt. Keyboard selection, however, does allow some selection correction in the form of moving the end location of the selection. This can be done on most systems using the ‘Shift’ and arrow

Referenties

GERELATEERDE DOCUMENTEN

“Voor de Rijksgebouwendienst was het erg wennen dat al die partijen mee gingen meebepalen,” (in: Alberts 2007). Waar de normale gang van.. 31 zaken zou zijn dat

Van de ‘verkiezing’ van de nieuwe EU president en de nieuwe Hoge Vertegenwoordiger kunnen we leren waar de echte macht ligt in de EU en vooral waar die niet ligt: niet bij de

De nieuwe economische geografie is belangrijk, met name voor de ontwikkeling van het nieuwe Europese en Nederlandse regionaal economische beleid in het landelijk gebied.

bubbles, rise veloeities and shape factors of the bubbles have been determined and compared with literature data. Same investigators suppose that the rise

licheniformis strains show lipolytic/esterase activity, the visual observation of lipid clearing on agar plates was not sufficient to establish which fatty acids

Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of

Het kan goed zijn om iemand te vragen zorg te dragen voor het doorgeven van informatie

Since the linear algebra approach links the algebraic variety V (I) to the nullspace of the matrix M δ , it should come as no surprise that the number of solutions of the