adaptable mobile user interfaces for e-learning ...

7 downloads 109752 Views 1MB Size Report
dynamically-generated Web content with calls to SOA services. .... fifth category of solutions provide mobile applications built for specific devices like iPhone ... Limited adaptation can be achieved by using specific add-ins and themes (Moodle,.
ADAPTABLE MOBILE USER INTERFACES FOR E-LEARNING REPOSITORIES Krzysztof Walczak, Jacek Chmielewski, Wojciech Wiza, Dariusz Rumiński, Grzegorz Skibiński Poznan University of Economics Department of Information Technology Mansfelda 4, 60-854 Poznań

ABSTRACT The paper describes a system enabling creation of adaptable mobile user interfaces for e-learning repositories, such as Moodle. The system design is based on a user interface adaptation framework, called ASIS, which enables combining dynamically-generated Web content with calls to SOA services. By using service calls as a means of interfacing elearning repositories instead of their standard Web interfaces, we can build new interfaces with significantly reduced complexity, focusing on important content and tasks, which is critical in case of small-size mobile devices. Furthermore, by utilizing full context information available at the access device, together with the device profile and a user profile, we can generate an interface specifically tailored for a given user, a given device and a given context of use. The user interfaces are dynamically generated based on interface templates coded in a high-level XML-based language specifically designed for this purpose. The paper describes the ASIS interface adaptation framework, its integration with the Moodle system and provides examples of mobile interfaces generated by the system. KEYWORDS E-learning, mobile devices, user interfaces, service-oriented architecture, ASIS, SOIL.

1. INTRODUCTION In the recent years, the use of e-learning platforms at schools and universities became increasingly popular. Latest surveys (2010) show that more than 90% of higher education establishments use some kind of e-learning system to support the education process (Greenberg and Zanetis, 2010). There is a variety of advanced e-learning platforms available, including commercial products and non-commercial open-source solutions. Examples of popular open source e-learning platforms include Moodle (Moodle, 2011), Sakai (Sakai, 2011) and ATutor (ATutor, 2011). These systems can offer great help for teachers and students in both teacher-driven and self-paced learning scenarios. Most importantly, learning with the use of such platforms may be performed in a classroom as well as in a distance learning configuration. The e-learning systems flourished together with the popularization of the Web – and later Web 2.0 – services and technologies. Consequently, most of these systems have been designed for use in the Web environment, exploiting Web browser capabilities to the maximum extent with the intention to provide users with the richest functionality and the best experience possible. However, at the present time, we can observe a quick change in the type of devices used for accessing the Web. Gartner predicts, that by 2013 the number of installed mobile devices, such as smartphones and browser-equipped enhanced phones, will exceed the number of installed PCs (Gartner, 2010). This trend is particularly visible in the case of young people, which constitute a vast majority of students, and who live in the “always on, always on-line” world and use not only a single desktop computer, but multiple devices like netbooks, tablet PCs, various types of PDAs and smartphones, like for example Toshiba TG01 (TG01, 2011), Apple iPod Touch (iPod, 2011) and HTC Touch Pro2 (Touch Pro2, 2011). Young people are used to be able to access any information, at any time, from any device, and they expect the same for educational content and applications. Even if mobile devices are in some aspects less convenient to use than a typical personal

computer, they have a big advantage – they can be used almost everywhere and at anytime: in a train, in a bus, in a cafeteria or in a university corridor. Enabling convenient access to e-learning repositories from hand-held devices would offer students additional opportunities for learning with the use of the available e-learning content. However, as compared to personal computers, mobile devices are usually much more limited in their capabilities, such as the communication bandwidth, the processing power as well as the offered presentation and interaction methods. Moreover, these devices significantly differ in their capabilities, making it very difficult to build a one-fits-all solution (Cellary et al., 2009). Therefore, it becomes necessary to adapt the interfaces and content for use on mobile devices. A learning content adaptation framework for mobile devices needs to meet several requirements: • support a wide range of target client platforms, at the same time maximizing interface and content adaptation to a particular device; • take into account full context information including interaction history and context data provided by the mobile device; • generate lightweight content to minimize bandwidth requirements as well as the necessary content processing and battery power; • maximize user interaction efficiency by providing easy access to selected high-level operations with simple user interface actions; • support a wide range of e-learning platforms. Existing solutions, as briefly summarized in Section 2, do not meet all the requirements listed above. For example, one of the important issues related to mobile access is the availability of context information. For PC environments context information is very limited, but for mobile devices the context information contains much more elements: user location (GPS or another location facility), usage scenario (sensor readings) and device capabilities. By exploiting the context information it is possible to offer a context-aware learning assistance (Lehsten et al., 2010) or to compensate to some extent limitations of the mobile devices. An important part of the context information are device capabilities. To generate an optimal mobile user interface for an e-learning application, a number of parameters describing device capabilities has to be taken into account. This includes the list of possible interaction channels and values of specific parameters describing these interaction channels, which limits the forms of communication that can be used by the final user interface. Apart from different device capabilities and other context information, efficiency of work with an application can be significantly improved by allowing users to adjust the application interface and choose the content and the functions, which they need most in given circumstances. Also, users may differ in their privileges to access different types of content. All these elements should be taken into account in the development of user interfaces for e-learning systems. In this paper, we describe a system called MILES – Mobile Interfaces for LEarning Systems. This system is based on a new approach to building adaptable user interfaces for e-learning repositories. In this approach, the e-learning repositories are accessed by the use of Web Services instead of typical Web based interfaces. Further, a new interface adaptation framework, called ASIS – Adaptable SOA Interface System, is used to generate user interfaces combining dynamically generated Web content and results of Web Service calls. By the use of the MILES system it is possible to provide students with a convenient and flexible way of accessing e-learning repositories. The way of accessing the content can be adjusted to the capabilities of the mobile devices used, to the users’ requirements, and to the current context (place, time, previous interactions). The adjustment is performed automatically, without explicit actions of the user. The system may support different e-learning platforms, providing centralized access to several distributed repositories. This paper is organized as follows. Section 2 contains an overview of the state of the art in adaptation of user interfaces, in particular for mobile learning. In Section 3, an overview of the ASIS framework is provided. In Section 4, the MILES system is described, including architecture and examples of MILES interfaces. Section 5 concludes the paper.

2. INTERFACE ADAPTATION The problem of dynamic creation and adaptation of user interfaces to content and applications has been widely studied (Browne et al., 1990; Al-bar and Wakeman, 2001; Chittaro, 2006; Wesson et al., 2010). Existing solutions can be broadly divided into five categories. The first category are declarative user interface languages, such as MXML (Meiser, 2010; MXML, 2011), XUL (Butter et al., 2007; XUL, 2011), XAML (Esch et al. 2007; XAML, 2011), UIML (Abrams et al., 1999; UIML, 2011) and UsiXML (Limbourg et al., 2005; UsiXML, 2011). Interfaces built with these languages can be highly sophisticated (Ye and Herbert, 2004), but – at the same time – such interface descriptions are usually complex and require complicated interpretation engines running on the client side. While this poses no significant difficulties on stationary PC platforms, practical implementations on mobile devices, such as PDAs and smartphones, are highly problematic due to the diversity of these platforms, low processing power available, and limitations of the communication channel. Therefore, most of the currently built mobile interfaces are based on HTML (HTML, 2011). The second category of solutions are HTML page generators such as PHP (Cao et al., 2006; PHP, 2011), JSP (Müller et al., 2005; JSP, 2011) and ColdFusion (ColdFusion, 2011). These tools can be efficiently used for generation of web pages and – in fact – are used in many practical applications on the Web (Boström et al., 2008; Moodbile, 2011). However, these solutions are usually limited to a single e-learning platform. Furthermore, providing support for a wide variety of devices and available context data would require very large amounts of code. The third category of systems simply use different versions of the interfaces for accessing with stationary and mobile devices (Yingling, 2006; MLE, 2011; MFM, 2011). This approach is very limited in its applicability, leads to inconsistencies and increases the amount of configuration that needs to be performed by end-users of these systems. In the fourth category of solutions, HTML interfaces are equipped with dynamic scripting features that enable their adaptation to the device capabilities on the client side. This solution, however, does not take into account full context information during the interface adaptation, significantly decreases bandwidth efficiency, and leads to security problems. Additionally, support for dynamic HTML features on mobile devices varies significantly. Finally, the fifth category of solutions provide mobile applications built for specific devices like iPhone (iPhone, 2011; Moodle4iPhones, 2011), Nokia (MLE, 2011; MOMO, 2011) and others (MHK, 2011; MoodleTouch, 2011).

3. THE ASIS FRAMEWORK 3.1. ASIS Architecture To enable building adaptable user interfaces for SOA-based applications, we have developed the ASIS framework (Chmielewski et al., 2010a, 2010b). ASIS – Adaptable SOA Interface System – is responsible for generating user interfaces for SOA applications, adapted to a particular user, the currently used access device and the current context. The ASIS framework creates the interface description on a request received from the end-user device, based on the available interface templates coded in a specially designed language, called Service-Oriented Interface Language (SOIL), parameters of the end-user device – either sent by the device or stored in a local database, and the current context of interface generation. The interface description is then sent back to the client device and rendered. At the client device, a typical Web browser or a specially designed application, extending a Web browser to provide more detailed context information, can be used. The overall architecture of the ASIS framework is presented in Fig. 1. The ASIS framework consists of four main elements: SOA Interface Generator Module (SIG), SOIL Interface Templates, Content Access and Adaptation Module (ADAM), and Content Upload and Adaptation Module (UMOD). The SOA Interface Generator (SIG) is responsible for structural adaptation of the user interfaces. SIG processes interface templates encoded in the SOIL language, generates the interface content

and sends the final content to the client devices. The SOIL Interface Templates enable SOA services to be presented to the end-users in a user-friendly way. In addition to the structural adaptation of interfaces, performed by the SIG module, ASIS also provides adaptation of particular interface components. This function is performed by the Content Access and Adaptation Module (ADAM). The module is responsible for providing access to multimedia objects and adjusting the formats and properties of these objects (type, format, resolution, sampling, precision, compression) to make them suitable for presentation on a particular stationary or mobile device. For example, to display a three-dimensional model of an object on a device, which is not 3D capable, a particular view of the object can be rendered to a 2D raster image. Formatted 2D text can be converted into an image to make the presentation independent of the text presentation capabilities of the particular mobile device.

Fig. 1. The overall architecture of the ASIS framework

The ASIS framework can be configured by the use of ASIS Configuration Application. It is a userfriendly graphical tool consisting of a number of specialized data managers for handling different types of data used in the interface generation process, such as interface templates, template instances, interface objects, device capabilities, context data, configuration settings and Web Services (Cellary et al., 2011).

3.2. The SOIL Language SOIL – Service-Oriented Interface Language – is a new XML-based language, which has been designed to enable efficient creation of user interfaces for SOA services and applications (Chmielewski et al., 2010a, 2010b). It provides commands that can control the process of interface generation and execute calls to SOA services. SOIL is independent of the content description language used (e.g., HTML, XML or textual PDF (Mars, 2011)). The SOIL language enables creation of interface templates combining static elements and dynamically generated elements (e.g., results of service calls). The SOIL language is divided into modules. Specialized SOIL modules, with functionality specific for a particular domain or application, can be added to the implementation. In the reference implementation, four SOIL modules are used: the Core Module (SOIL CM), which contains basic language elements independent of a particular application; the Database Module (SOIL DB), which contains commands that enable retrieving and updating data in databases; the Service Module (SOIL SRV), which contains elements enabling execution of requests to SOA services implemented – for example – as SOAP services, and the Application Module (SOIL APP), which contains elements specific to a particular application. For example, the Application Module could be used to introduce commands retrieving Web Service user interface cues (for example, provided according to WSRP specification (Chen et al., 2005; WSRP, 2011)) and use them to build the final user interface.

4. MILES – MOBILE ACCESS TO LEARNING REPOSITORIES 4.1. Connecting ASIS to LCMS For developing the MILES system, we selected Moodle (Moodle, 2011) as the base LCMS. Advantages include extensive functionality, extensibility and high popularity of the system in schools. In the Moodle system, users can get access to the course material using a Web browser, but the interface is hardly readable on small size devices. Limited adaptation can be achieved by using specific add-ins and themes (Moodle, 2011), but the level of adaptation is not sufficient to meet the requirements described above. To enable visualization of the course content by the ASIS framework, we have developed a collection of Web Services that enable accessing the content available within the Moodle LCMS. The overall architecture of the MILES system is presented in Fig. 2. The system consists of a Moodle instance and the ASIS framework connected by the use of the Web Service interface.

Fig. 2. MILES: Moodle LCMS and the ASIS framework

To enable the ASIS framework to invoke Moodle Web Services, the SOAP protocol was chosen. This protocol is supported by the Zend SOAP (Zend SOAP, 2011) server integrated in Moodle. To be able to call defined services, first the configuration of the Moodle system has to be adjusted by switching on the Web Services support. The next step is to install the class file with methods, and add them to the list of available Web Services, that can be called from the external systems. All Web Service methods have been implemented in the PHP language. The following methods form the basic interface: • get_user_by_login – the method gets user login as the input and as returns user ID as the result; • get_user_courses – the method gets the ID of the user as the input parameter and returns the list of user’s courses. This list contains the course names and ID numbers; • get_course_sections – based on a given course ID the method returns the number of sections; • get_course_resources – the method returns all resources available within a given course and section. Specification of each resource consists of a name and the type of the resource and a direct link from where the content can be retrieved. These methods are sufficient to enable users to display their courses and associated resources. To provide maximum flexibility in configuring Web Services used for connecting to the LCMS systems, a special Service Configuration Manager tool has been developed as part of the ASIS Configuration Application (Fig. 3). The tool enables configuring service parameters and service parameter values of external services used by the SOIL interface templates. By the use of the tool an administrator can create, delete and copy different service configurations (sets of parameters), thus enabling connections to different LCMSs. Then, by the use of another tool – the Service Configuration Sets Manager – the administrator can switch between the available preconfigured sets of services. The Services Configuration Manager window is split into two panes (Fig. 3). On the left side, there is a list of service configurations together with descriptions. There are four services configured in this example, as specified above. Each configuration is associated with a list of parameters, which is visible on the right side of the window. In the example provided, the parameters associated with get_user_login service are displayed: operation, WSDL, namespace and the required service call parameters.

Fig. 3. ASIS Configuration Application – Service Configuration Manager

4.2. Adaptable User Interfaces in MILES Access to the learning content in typical LCMSs is realized by the use of a Web browser and collections of dynamically generated Web pages. In most cases (and also in the Moodle LCMS) the content is generated on server side with the use of the PHP language. A typical page presenting the list of user’s courses is presented in Fig. 4. High resolution of typically used computer screens guarantees easy and clear presentation of the content and smooth interaction. This, however, is not the case when such Web pages are accessed from a mobile device (Fig. 5). Large number of elements on such a page makes it unreadable on a small screen of a mobile device. Even if zooming capabilities are available, only a part of the zoomed page is visible. It reduces user's awareness of the content. Moreover, accessing different interaction elements scattered around the page requires a lot of additional interaction – scrolling, shifting, etc.

Fig. 4. List of available courses presented in a regular Web browser

A page presented in a regular browser may contain a lot of different elements (including images, movies, graphical design of a page) that are not necessarily related to the essential tasks to be performed by a user when accessing the platform from a mobile device. An interface presented on a mobile device should be equipped only with the most important elements allowing users to perform the most important tasks – mainly accessing the course content. At the same time, an interface should be adjusted to the context – in particular the capabilities of the device used.

Fig. 5. Typical list of available courses presented in a Web browser of a mobile device

With the structural and content adaptation capabilities of the ASIS framework, it is easy to build in MILES simplified user interfaces for the use on a mobile device. Moreover, as opposed to the available mobile LCMS interfaces, the MILES interfaces can be adjusted to particular device capabilities, user preferences and current context. An example of such an interface, cleaned from all unnecessary elements, but permitting users to access their courses and then all course objects is presented in Fig. 6. The interface presented in Fig. 6a displays a list of user courses, while the interface presented in Fig. 6b displays objects within a selected course. All necessary data, like the list of courses, the list of topics, objects belonging to the course, are retrieved through the Web Service interface described in Section 4.1. As an example, a fragment of code in the SOIL programming language, responsible for extraction of courses is presented below:

Suggest Documents