1. INTRODUCTION

A software application is the product of completion of many phases such as planning, analysis, requirements, design, development, testing and deployment. The development phase which includes design, actual coding, and testing has own importance in the software development life cycle (^{Kantorowitz & Lyakas, 2005}). Software design is defined as the interface between the analyzed requirements and actual software development.

E-Learning and M-Learning are technology based learning systems. M-Learning is defined as learning using mobile devices ^{(Fermoso, Mateos, Beato, & Berjón, 2015)}. M-Learning provides flexibility of learning at anytime, anywhere, and by anyone ^{(Khan, Al-Khanjari, & Sarrab, 2016; Sarrab, Elgamel, & Aldabbas, 2012}; ^{Sarrab, Elbasir, & Alnaeli, 2016}).

The proliferation of mobile devices such as PDA, smart phones has created opportunities as well challenges in design and development of application ^{(Ogbuji, Onuoha, & Izogo, 2012)}. The GPS navigation and location tracking are used for learning during museums visits ^{(Duran-Limon, Siller, Hernandez-Ochoa, Quevedo, & Robles, 2014; Iqbal-Faruque, Aisyah-Husni, Ikbal-Hossain, Tariqul-Islam, & Misran, 2014)}. The accessibility of Internet based mobile applications is affected due to network ^{(Kandadai, Sridharan, Parvathy, & Pitchaimuthu, 2017)}

Design is an important aspect in software development ^{(Chu & Lo, 2013)} but the content based application such as M-Learning leads to the complexity in design phase. The arrival of multimedia applications has further increased the complexity ^{(Chen, L. et al., 2013)}. The learning domain are facing design and development challenges due to evolutionary nature ^{(Páivárinta & Smolander, 2015)}.

This research proposes an integrated design and development model for M-Learning application development. The model has two parts: pedagogy and software application. The pedagogical design is a systematic approach which uses methods, models, theories in order to enhance the learning outcome (^{Conde, García-Peñalvo, Alier, Mayol, & Fernández-Llamas, 2014}). The pedagogical design defines plan, scenario, format, content, structure, environment, delivery system, and implementation strategy ^{(Romiszowski, 2016)}. The pedagogical design contains instructional and learning design. The learning design is concerned with designing the learning content using different styles of learning such as visuals, aural, read/ write, kinesthetic ^{(Reigeluth. 2013)}. Learning design is related to students for tasks/ opportunity such as lesson, a unit of learning or a course (^{Khanghah & Halili, 2015}; ^{Soflano, Connolly, & Hainey. 2015; Yang, Huang, Gao, & Liu, 2014)}. Instructional design is related to instructors for systematic management of the systematic management of learning process and learning plan such as lecture format. Learning theories are Learning theories are descriptive and make general statements about how people learn. For example, Constructivism, Cognitive and Behaviorism. The software application design consists of architectural level design, high level design, detailed design, and usability design. Architectural design defines the overall structure of the system or module and relationship between them ^{(Clark & Mayer, 2011)}.

The proposed model presents a structured and integrated approach for utilizing technology, context, usability, and pedagogy in the design and development of M-Learning content and application. The model involves identification of stakeholders, modeling and classification of requirements, design of learning content and software, creation of test cases, development of learning content and software, finally testing and evaluation.

The paper is structured as follows. There are seven sections. Section 2 is about the methodology adopted for the research. Section 3 relates to the literature review and existing problems. Section 4 describes the proposed model. Section 5 involves the discussion. Section 6 illustrates the feasibility of the model, and the last section summarizes the findings.

2. METHODOLOGY

This study has used literature reviews in order to explore the possible pedagogical and application design solution for M-Learning. The steps used in the process are shown in Figure 1. The first step defines the problem, followed by discovering, filtration, and finally reporting. The digital databases explored for problem analysis, relevancy, and possible solutions includes: Science Direct, IEEE xplore, Scopus, and Springer using simple and advanced search. Searching has been conducted individual as well as in combination of search terms. The search terms include: M-Learning application design, learning design, M-Learning design, model etc. Boolean Operators such as AND, OR, *, has been used to retrieve articles from databases. For example, the keywords include "M-Learning design", "learning design", and "model", and application of Boolean search is as follows: (model* OR approach*OR method*OR process*) AND (M-Learning application design * OR learning design *OR M-Learning design*). Although the searching leads to a number of articles however, articles published between 2005 and 2016 with relevance to M-Learning application design have been studied thoroughly.

Fig. 1 Methodology followed in the research 

3. LITERATURE REVIEW AND PROBLEM IDENTIFICATION

The research in mobile application development is still theoretical and limited in scope ^{(Kim, 2015)}. M-Learning is a type of mobile application which exerts considerable influence on education ^{(Muyinda, 2007)} but designing M-Learning is a complicated process. M-Learning design needs attention ^{(Alden, 2013)} due to contextual nature of mobile applications and limitations of mobile devices (e.g., screen size, low processing power, and energy limitation). Software engineering has numbers of design methods ^{(Dañado & Paterno, 2014; Marinho et al., 2013)} however these methods cannot be applied to M-Learning due to uniqueness ^{(Zhu, 2005)}. For instance, E-Learning design models ADDIE (Analysis, Design, Development, Implement, and Evaluation) ^{(Chittaro, 2011; Naismith. Lonsdale, Vavoula, & Sharples, 2004)}, ASSURE (Analyze learners, State objectives, Select media and material, Utilize media and materials, Require learner participation, Evalúate and revise) ^{(Seels & Richey, 2012)}.

M-Learning materials require a detailed and systematic planning of design and development process ^{(Heinich, Molenda, & Russell, 1989)}. There are technological and pedagogical challenges such as user interface design, múltiple devices ^{(Mahazir, Norazah. Rosseni, Arif, & Ridzwan, 2015)} variation in networks (wireless home área networks (HAN) and wireless wide área networks (WAN), business área networks(BANs) ) ^{(Develi & Kabalci, 2017)}. The designing of course bring challenges to the course designers ^{(Wu, Lee, Chang, & Liang, 2012)}. Integration of infrastructure and application is another challenge in ubiquitous application ^{(Stanton & Ophoff, 2014)}. The summarized output of extensive analysis of the articles on M-Learning design is shown in Table 1 and 2.

Pedagogical design should be taken care while developing learning applications such as understanding different learning style (Visual, Verbal, and Inductive etc.) ^{(Parsons, Ryu, & Cranshaw, 2007)}. Additionally, the design phase if not handled properly would affect the M-Learning development and implementation ^{(Franzoni-Velázquez, Cervantes-Pérez, & Assar, 2012)}. It has been observed that there are researchers who are concerned only about instructional design of M-Learning. Moreover, as far as the M-Learning is concerned the existing design methods do not address the design of both pedagogy and software application in parallel which often results in compatibility and integration issue. Using the existing approaches design become complex and fail to fulfill the objectives resulting adverse effeets on learners ^{(Hsu, Ching, & Snelson, 2014; Zhu, 2005)}.

4. PROPOSED MODEL

The proposed design model has two important components: pedagogy and software application. The pedagogy defines the educational part while the software application defines the actual application supposed to run on mobile devices. The design model is break down into modules: a. Requirements modeling and identification of stakeholders b. Design of learning content, software application, and test cases c. Development and testing of learning content and software application e. Evaluation and feedbacks. Figure 2 illustrate the proposed integrated design model for M-Learning software application and pedagogy.

Fig. 2 Integrated Desing Model for M-Learning Pedagogy and Application 

5. DISCUSSION

The design phase of M-Learning software development life cycle affects the subsequent phases such as development, testing, and deployment. Therefore, it's necessary to consider all the constraints and criteria while designing the pedagogy and software application for M-Learning. The proposed design approach has considered the limitations of the existing approaches by advocating integrated design model. These limitations would be discussed one by one.

First of all, the proposed model has considered the pedagogical design and application design as a one task with subtasks undergoing in parallel. The common design issues in M-Learning is communication and interactivfty between different components ^{(Alvarez, Alarcon, & Nussbaum, 2011)}.

Likewise, several researchers performed case study in M-Learning and concluded that the instructional design should be aligned with the objectives and technology ^{(Capretz, Ali, & Ouda, 2012; López, Royo, Laborda, & Calvo, 2009; Naismith & Corlett, 2006)}. Not all kind of learning material can be uploaded in an application ^{(Chou, Block, & Jesness, 2012)} . In an article ^{(Cheon, Lee, Crooks, & Song, 2012; Williams, 2009)} defined a design approach for Augmented Reality with emphasis on pedagogy and software application together.

Starting with software application design, this design model has considered three levéis: architectural, high level and detailed. For instance ^{(Chen, D., Chen, Huang, & Hsu, 2013)} claimed that software design is composed of architectural design, interface design, detail design. At the architectural level design, the software designer identifies the design solution of the problem. The sub systems, constraints, interacting interfaces, and architectural styles (Batch sequential, Client-server, communicating processes, Event systems, Object-oriented etc.) are explored. At the high level design breaking of architectural level design into modules is performed. The sub systems are now represented as modules. During the detailed design level module details (data structures and algorithms) are designed which would be further used in the development phase.

As far as pedagogical design is concerned the proposed model has addressed instructional design and learning design. Most of the M-Learning design and development fails to consider these two important áreas. M-Learning software development suffers from good pedagogical designs ^{(Zhu, 2005)}. M-Learning software development process should follow clearly defined pedagogical design principies depending on the learner type, context of use, and interactivity ^{(Laru, Naykki, & Jarvela, 2015; Looi, Sun, & Xie, 2015)}.

In the proposed model the pedagogical design starts with the instructional design which is equivalent to architectural design of software application. ^{(Lehner & Nósekabel, 2002)} after reviewing M-Learning articles from 2007 to 2014 on the design of science apps pointed out that instructional principies are rarely implemented during the design and suggested that instructional designers and researchers should pay considerable attention. Instructional design and development should incorpórate pedagogical principies and theories ^{(Zydney & Warner, 2016)} specially proposed for M-Learning ^{(Lim. Oakley, & Liu, 2013)}. Instructional materials should be presented in granular fashion ^{(Huang, Lin, & Cheng, 2010)} and use of micro lecture format ^{(Lowenthal, 2010)}. Designing digital contents for M-Learning require analyzing cognitive skills of students ^{(Gu, Gu, & Laffey. 2011)} and presenting learning content logically ^{(Chiu & Churchill, 2014)}.

Next step is learning design. The proposed model dictates to consider the learning styles and learning theories such as behaviorist (instruction oriented), constructivist (construction of knowledge or task-based activfties), collaborative (group work, dialogue, and discussion) in learning design. It is important to consider learning styles and learning theories ^{(Yang, Li, & Lu, 2015)}. The learning design is equivalent to high level design of software application. The detailed designs have to be in congruence with the learning design. For any discrepancy necessary changes have to be included into practice. There should be alignment of learning content, learning context, and the adjustment of instructional strategies ^{(Schwabe & Góth, 2005)}. The design should be tested to make sure that design and instruction principies are implemented ^{(Tsai, Young, & Liang, 2005)}. A pedagogical design should adopt a theory of learning Mayes (2004). Learning theory, instructional design theory, and instructional design are interlinked ^{(Zydney & Warner, 2016)}.

The instructional as well as learning design job is performed by Instructional and Learning designer. The common tasks are: look and feel, organization and functionality of instruction, blueprint for the course, outline of subjects etc.

The last important component of the proposed model is testing. The testing of product is important for acceptance . There are three types of testing performed in the proposed model: emulator, cloud, and crowd sourced. The test cases are generated for validation of the application at stages during high-level design. Even though the development of application starts with requirements elicitation however including the testing processes from very early stage is required ^{(Moraes, Andrade, & Machado, 2016)}. Software testing is an important step in development process for building a reliable and correct software ^{(Belli, Budnik, Hollmann, Tuglular, & Wong, 2016)}. Henee in the proposed model once the application has been tested well it's accepted for use.

6. COMPARISONS WITH ISO/IEC 12207:2008 AND ISO/IEC 19796-1 STANDARDS, EXPERTS REVIEW, AND RESULTS

The existing design approaches for M-Learning pedagogy and software application from Table 2 (coded as A, B, C, D, E, F, G, H, I, J, K, L, M and the newly proposed approach is coded as N) is compared with ISO/IEC 12207:2008 and ISO/IEC 19796-1 standards. These comparisons are shown in the Table 3 and Figure 3. Out of seven process groups in ISO/IEC 12207:2008 standard five have been taken for comparisons. The comparison with ISO/IEC 19796-1 covers five processes out of seven. ISO/IEC 19796-1 is a standard that help educational organizations to develop quality systems and improve the quality of their processes, producís, and services ^{(Pawlowski, 2007)}.

Fig. 3 Comparison of existing M-Learning desing approaches with ISO/IEC 12207:2008 and ISO/IEC 19796-1 

The comparison of existing design and development approaches and proposed Integrated Design Model with ISO/IEC 12207:2008 and ISO/IEC 19796-1 (Table 3 and have identified the stakeholders and their roles during Figure 4) shows that all the five criteria are not satisfied design, development, and testing steps. In the proposed by any of them except "N". Also, most of them considered integration of learning content and software application, evaluation, and testing steps. None of them application modules have been emphasized.

Fig. 4 SWOT analysis of the proposed Integrated Design Model for M-Learning Pedagogy and Application. 

Experts Review

IT experts examined the the proposed design model from April 2017 to July 2017 to provide feedbacks on the suitability in the context of learning. Out of twelve experts 41% have 2-3 years and 25% have more than four years' experience in M-Learning design and development.

The experts have positive perception towards stakeholders' identifications, requirements classification and modeling before actual design starts. Eight of them view it's important to go for parallel design and development of learning content and software while two of them disagree. When the reasons for their disagreement were investigated it's found that the disagreement is due to less experience in M-Learning. The experts have positive opinión for including the learning styles and learning theories during instructional and learning design as shown in Table 4. This further supports the pedagogical design steps in the proposed model.

The experts are of the view that the end users should be involved in testing and testing should be performed at both learning content, and learning software which supports the proposed model steps of testing. Table 5 shows the experts feedbacks on testing.

Lastly, the experts were asked an open ended question to provide feedback on the overall integrated design model. Out of twelve, nine of them accepted the steps of the design model. The experts have found design, development, and testing steps appropriate and acceptable.

The SWOT (Strengths, Weaknesses, Threats, and Opportunities) analysis of the proposed model is shown in Figure 4.

7. CONCLUSION

Successful implementation of M-learning is influenced by design approach followed for pedagogy and application. The proposed model is based on the critical analysis of the existing literatures. The proposed design model is the first of its kind where the designing of both pedagogy and application takes place in parallel. The proposed model emphasizes requirements classification and modelling, identification of stakeholders, pedagogical design and development, application design and development, designing test cases, and testing of the developed application.

The requirements are classified into functional, non-functional, and pedagogical. The stakeholders are categorized as software application and pedagogical stakeholders. The model considers three level of application design: architectural, high, and detailed design level. The pedagogical design consists of instructional design and learning design. The pedagogical stakeholders are involved in pedagogical design.

The test cases are designed during software application and pedagogical design. The test cases are used to find out compatibility between learning and application design, suitability of learning content and application, and alignment of the developed M-Learning application with the elicited requirements.

Lastly, the proposed model has been verified by comparing with ISO/IEC 12207:2008 and ISO/IEC 19796 1 standards and IT experts review. The feedback is positive in both cases. In future M-Learning application would be developed following the steps of proposed model and application design principies would be proposed

CONFLICT OF INTEREST

The authors have no confliets of interest to declare.