Innovative learning and teaching scenarios in Virtual Campus Mirko Cesarini
[email protected]
Sam Guinea Montalvo
[email protected]
Licia Sbattella
[email protected]
Roberto Tedesco
[email protected]
Dipartimento di Elettronica e Informazione Politecnico di Milano Milan, Italy
Abstract: The paper discusses the advantages of Virtual Campus, a platform integrating authoring, fruition and validation modules, where flexibility and personalization of the whole system are main goals. To ease the authoring activities, Virtual Campus leverages both the reusability of existing courses - while building new ones - and a customization process that allows to adapt an existing course to different situations. To achieve course and material content reusability a model has been developed that - with the aid of metadata - permits to store, classify, and browse didactical materials, making them available for flexible learning paths. A run-time infrastructure has been developed that can drive a student along a learning path, previously designed by the teacher. A workflow management system manages every student’s state of advance, allowing path customization. Finally a set of monitoring tools assist the students during their learning activities providing both suggestions to the students and feed-back to the teachers. Isolated and cooperative work sessions, tutored and untutored participations, virtual and “real” presence in an extended or restricted study group, synchronous and asynchronous communications have been studied for testing the global system.
1. Introduction The Virtual Campus[1] platform (see Mainetti et al. 2002) is an open system for the design, deployment, fruition, and evaluation of reusable learning materials realized at Politecnico di Milano. Its main objectives are: to support design, composition, and reuse of learning material; to support fruition of learning material in individual and cooperative usage; to support analysis of students’ behavior (learning, relational, and normative aspects); to analyze power consumption in all envisaged scenario; to experiment the usage of the learning platform and of advanced learning objects in Computer Science (CS) courses[2]. Reusability and flexibility mainly characterize the Virtual Campus Platform. A rich conceptual model and a modular architecture, globally supported by a workflow engine, provide integrated interfaces for authoring, fruition, tutoring and validation. The paper is going to present the scenarios of interaction devoted to the teachers, the students and the tutors, emphasizing: the aggregation mechanism which allows both to create complex learning objects from atomic ones and to define the same set of metadata at different levels of granularity, the three-level conceptual architecture which allows the definition, the suggestion and the choice of different learning pathways, the use of metadata and of dynamically detected use-data which allow the definition of different profiles (the ones related to students’ behaviour but also the ones related to the quality of learning objects and to the agreement of teaching and learning scenarios).
2. Platforms for e-learning In the e-learning domain several research projects and commercial applications are available. CentraOne (see Centra 2003) is a collaboration and communication system and focuses mainly on the sharing of information among learners and on enabling collaborative creation of artifacts.
[1] [2]
The Virtual Campus Project is sponsored by Microsoft Research (UK). More general information on the platform at http://www.elet.polimi.it/vcampus.
BlackBoard Learning System (see BlackBoard 2003) is an example of a Learning Management System and aims at providing a complete e-learning environment supporting authoring, management, and fruition of courses. All aforementioned systems fail to give to teachers the opportunity to define a true course structure, therefore it is not possible to define precise “course paths” through the teaching material. MediBook (see Steinacker et al. 2001) is a relationship-based system and allows teachers to define a course structure by means of logic relationships among the course components. Flex-eL (see Lin et al. 2001) is a workflow-based system and allows teachers to define a course structure as a workflow. Flex-eL provides a process-modelling tool to capture the learning process and view it as a stream of activities. Workflow-based systems define a strict temporal order between activities. In some cases this forces teachers to impose unnecessary constraints. CoLab (see Hoyos-Rivera et al. 2002) is a rule-based system and customizes the fruition of LOs by mean of rules, giving teachers the opportunity to define “learning paths” through the teaching material. Rule-based systems provide support to the definition of constraints on fruition paths, but do not offer mechanisms to let teachers define precise paths if this is needed. Finally, Edutella (see Dolog et al. 2003) is an interesting example of a peer-to-peer learning system. We argue that the mentioned approaches provide solutions that address only partially the overall problem of supporting the creation of learning objects fruition paths.
3. The Conceptual Model The core concept in the Virtual Campus conceptual model is that of Learning Object (LO). A learning object is anything that can be used to convey some concept from a teacher to a student. Structurally a LO is composed of metadata and content: • Metadata are descriptive information about resources for the purpose of finding, managing, and using them more effectively (see Fischer et al. 2001). For content providers or publishers, metadata simplifies the discovery and access to their resources so they can reuse them. • Content is the actual didactical information a student is allowed to exploit. Metadata for Virtual Campus have been specified according to an extension of Learning Object Metadata (LOM), a standard defined by the Learning Technology Standardization Committee of the IEEE (see IEEE Learning Object Metadata Working Group 2002). A LOM metadata instance for a LO describes relevant characteristics of the LO to which it applies. Such a description is organized in several categories including general information on the resource, history and status information, technical requirements and characteristics, educational attributes, intellectual property rights, and use conditions. Our extension, called VC LOM, mainly concerns metadata of the educational category and focuses especially on relationships between LOs, preconditions, learning objectives and interaction modalities among students, teachers and the platform. VC LOM modifies the standard LOM relationships, giving them a precise semantics. A standard set of relationships is able to express “requirements” (LO A must be learned before LO B), “equivalence” (LO A gives the same information as LO B), “optionality” (LO A is an optional in-depth study of LO B), etc. Such a description expresses constraints to the fruition and is exploited to generate all the compliant pathways through LOs. VC LOM metadata support several interaction modalities in order to make use of diverse LO contents and allow cooperation among students and between students and teachers. In particular, LO fruition can either be individual or require a group participation (if a group is required, LO metadata must specify the related cardinality); LO collective fruition can either be “synchronous” (all participants must be on-line at the same time) or “asynchronous”; LO fruition can either be “auditable” by other students or “not auditable”; LO fruition can require the presence of one or more “guides” (tutors). Other metadata allow to specify particular conditions that must be verified in order to permit the fruition of a given LO (as an example, administrative requirements), and the skills obtained after a LO fruition. Such attributes specify preconditions and learning objectives and define how the fruition of LOs can modify a user profile. In the Virtual Campus platform a lesson, an exam, even an entire course is modeled in terms of LOs. Thus, LOs can either be elementary items (Atomic Learning Objects) or they can be defined as the aggregation of other LOs (Complex Learning Objects). While Atomic LOs may represent single lessons or learning units and be associated to specific contents such as slides, videos, tests, homework, etc., Complex LOs represent parts of courses or entire courses. Being composed of various LOs (either Atomic LOs or other Complex LOs), they establish particular
relationships between component LOs. At fruition time, these relationships may force the order in which students exploit such LOs. LOs can be refined and customized by teachers before making them available for fruition. Within Virtual Campus several different actors can be distinguished: Learners, Organizers, Authors, Teachers, and Tutors. Learners use LOs with the intent to transform information into knowledge. Authors create Atomic LOs by providing course material (contents) and associating specific metadata values to them. Organizers create Complex LOs (itineraries of LOs) by assembling LOs and associating specific metadata values to them. Teachers customize and deliver existing LOs to their own classes. Finally, Tutors support Teachers and Learners in their activities. All authoring activities are performed at three main levels of abstraction: Starting from an abstract (and highly reusable) logical representation, a LO undergoes a customization process whose goal is to best fit the Teachers’ specific requirements. During such a process, three steps can be identified, each of them increasingly adapting the course representation to Teachers needs: • The Reusable Level where general-purpose Reusable LOs are defined. • The Didactical Level where workflows (called Didactical-level Complex LOs) are generated starting from the constraints defined at the higher level. • The Fruition Level where additional details are added, thus leading to the definition of Fruition-level Complex LOs and Fruition-level Atomic LOs.
4. The overall architecture The Virtual Campus platform is composed of two main subsystems (Fig. 1): the Authoring Environment and the Fruition Environment. Consistently to the authoring activities and the levels listed above, the Authoring Environment provides editors with the means to define Reusable Atomic LOs and Reusable Complex LOs, a Didactical-level Complex LO generator/editor that automatically provides a first version of the workflow associated to a Complex LO and then supports Teachers in customizing it, and, finally, a Fruition-level LO tailoring tool that supports the insertion of all fruition-level details such as timetables, classroom composition, etc. LOs produced within the authoring environment are then serialized into a fruition package, based on SCORM 1.2 (see Advanced Distributed Learning 2001). The SCORM standard is extended providing information regarding the structure of Complex LOs, both in terms of the Reusable Level definitions and as a workflow. The former is a higher level description that allows Organizers to easily reuse learning material; the latter permits Teachers to precisely define the didactic pathways available to Learners at fruition time. The Fruition Environment enables navigation of these didactic pathways by Learners. Such a fruition can be guided by Teachers, and can involve either individuals or groups of Learners. A workflow engine[1] enacts the fruition, guiding Learners and Teachers in the execution of the activities related to the usage of the LO. The engine permits to manage independently the state of every student according to the course specification done by the Teacher. Moreover, the engine ensures every student accesses LOs in the correct order. The fruition environment includes tools used to exploit all contents associated to LOs. These tools, other than standard productivity tools such as Microsoft PowerPoint or Acrobat Reader and tools supporting synchronous collaboration such as Microsoft NetMeeting, include tools that have been specifically developed within the Virtual Campus project (Lezi.NET, WebTalk, PeerVerSy, and WebBoardExport).
[1]
Workflow is concerned with the automation of procedures where documents, information or tasks are passed between participants according to a defined set of rules to achieve, or contribute to, an overall business goal (see WfMC 1999).
Reusable LO editor Author
Organizer
Didactical-level Complex LO generator & editor Fruition-level LO tailoring tool
Raw Contents Teacher
Reusable LOs
Learner Lezi
WebTalk
Teacher
Tutor
WebBoardExport
PeerVerSy
User interaction environment Didacticallevel LOs Fruitionlevel LOs
Tutoring & Validation Module
Instrumented Fruition Engine
Workflow engine
Fruition DB
data
Profile DB
LO authoring environment
LO fruition environment Extended SCORM
Extended SCORM
Figure 1: Virtual Campus - overall architecture.
5. The Authoring Environment The Virtual Campus Authoring Environment was envisioned to support the work of LO Authors, Organizers, and Teachers, to assist and guide them during the process of creation and modification of reusable LOs. LO reusability is, in ultimate analysis, one of the principal goals in our project. It empowers Teachers, allowing them for a more flexible and efficient way of creating Complex LOs for their didactical purposes. Complex LOs can in fact be created by Organizers through the composition of reusable “building blocks”, which are other LOs contained within the Virtual Campus Repository. In the project, reusability has been achieved by organizing a LO’s life-cycle into three different phases. In fact, inside the Virtual Campus Architecture a LO can be treated at a Reusable Level, at a Didactical Level and at a Fruition Level. The authoring environment is the tool that aids us during the development of LOs along the first two levels. It has been developed following a client-server paradigm. The client is a stand-alone application that, by use of XML and a series of Web Services published by the server, can interact with the Virtual Campus Repository. The application is provided as an extended and personalized version of MS Visio, to take full advantage of its drawing capabilities. At a Reusable Level the tool helps create reusable “building blocks” which can vary from fairly simple LOs to very complex and articulate compositions of LOs. Inside the tool an Author can create Atomic LOs, describe them with metadata and insert them into the Virtual Campus Repository in a matter of seconds and if errors are made, the metadata can always be modified in a second moment. The creation of Complex LOs is equally easy. First of all an Organizer must create a new Complex LO describing it with a subset of compulsory metadata. This new LO immediately becomes a container for eventual reusable “building blocks”. In fact Organizers can now search for LOs within the repository by use of metadata or by viewing them in alphabetical order. Once one suiting their needs has been found, it can be inserted into the new Complex LO. If no satisfying LO exists a new one can eventually be created on the fly, described by metadata and inserted, both into the repository and into the new Complex LO. When all the necessary LOs have been inserted they must be organized using the four logical relationships provided by Virtual Campus: IsRequiredBy, IsAlternativeTo, References and RequiresOnFailure[1]. The tool will help and guide in their appliance, preventing the user from drifting from the compositional guidelines specified in the Virtual Campus Conceptual Model. By using these logical relationships no real order is given but constraints are established with the intent of determining allowable didactical paths. The focus is clearly set on didactical relationships between LOs and the result is an easily interpretable graphical model (Fig. 2). The Didactical Level was introduced in our three phase LO life-cycle plan to be able to give Teachers the opportunity to customize work previously done at the Reusable Level. Inside the Authoring Environment the tool automatically converts Reusable Level LOs into Didactical Level LOs, representing them in a workflow-like manner. [1]
A complete study of the logical relationships provided by the Virtual Campus Conceptual Model goes beyond the scope of this paper. Please refer to the Virtual Campus Website for more information.
Figure 2: The Virtual Campus Authoring Environment. On the left an example of a Reusable Level LO definition is shown. On the right, its automatically generated Didactical Level definition has been greatly customized. In this case the teacher has decided to generate a linear path. The reason for this conversion is to provide a visual representation that could help the viewer focus on explicit paths. By introducing the Didactical Level a new kind of reusability is ensured: multiple copies of Didactical Level LOs can be derived from a single Reusable Level definition, each with different didactic path specifications. Working directly in the visual environment the Teacher can proceed to customize these paths by literally rearranging them. The tool will obviously not permit the breaking of logical constraints established at the Reusable Level (e.g. no additional LOs can be added to the workflow paths) and will guide the Teacher throughout the entire process. Typical customizations can be: the choice of a specific LO in a situation where alternatives are presented or the creation of a single path in a given multi-parallel path situation. The result can be a simple linear path (as in the example in figure 2) or a more complex set of pathways, all available to Learners who can then choose on their own at fruition time. Once a didactical LO has been finalized it can be inserted into the Virtual Campus Repository. At the Fruition Level, Teachers enrich LOs with administrative information (the list of enrolled students, timetables, etc) providing ready-to-be-used LOs to Learners. Once again, multiple copies of Fruition Level LOs can be derived from a single Didactical Level definition, each with different administrative-related information.
6 The Fruition Environment The current interface students use to interact with the system is web based. Although other possibilities are under evaluation, this seems actually the best choice to ensure the availability of the application in all the heterogeneous computing environments that a student has to interact with. In fact, a student should be able to access the application from the campus labs, from library desktops and from home, switching between them many times during a day. The web browser is required to display only standard HTML and multimedia contents. There is no limitations to multimedia formats that can be used as teaching material, the only requirement is that a standard web browser should be able to display it (either natively or with the aid of a plug-in) or download it for off-line viewing. A student wishing to use the system connects to an URL and receives a web page with a form that should be filled with a username and a password. After the login procedure has succeeded, a web page containing the so called “working area” is presented. For most of the time the student will interact with the working area. It is divided in two frames: the left one containing the history of visited LOs; the central one used to show material about the LO the student is required to study. The working area right frame can vary noticeably according to the LO content type provided. (Fig. 3 and Fig. 4). All the web pages are dynamically generated server side, in this way the computation skills required to the web browser are very low. This approach also allows to meet another requirement: all the LOs building a course should not be freely available to a student. Instead, they should be presented to her/him one after another in an ordered way, accordingly to what stated by a Teacher during the authoring phase and also to the Learner’s choices during fruition. The richness of our model allows a Teacher to specify a course where multiple learning paths can be depicted. In this way a student has some degree of freedom in choosing how to build her/his learning path.
Figure 3: WebTalk is a collaborative 3D environment providing chat facilities. The right frame of the working area is divided into two spaces containing a 3D interface through which the participants can explore the virtual environment and chat facilities.
Figure 4: Video and slides. In this case the right frame of the working area is divided into four spaces containing: a streaming video, synchronized slides, an index to the video and in the lower-right corner various reference material.
Consequently the sequence of LOs to be presented to a student also depends on her/his choices. In fact, when the course definition requires the student to express a preference between two or more LOs, or to eventually study an optional one, a choice is presented. The student can communicate her/his decision by selecting one of the proposed alternatives, according to which the next LO will be presented. The history contains a list of the visited LOs, and is useful when the students need to examine a previously accessed LO, without modifying her/his current advancement state. The workflow management system is the key concept of the fruition environment in the architecture since it maintains a state for each student, therefore permitting independent paths. The web-server is a bridge between the student web browser and both the workflow engine (that handles the knowledge on which is the next LO to be presented to a student) and the repository containing the LOs.
7 The Tutoring and Validation Module The Virtual Campus fruition engine is instrumented with a monitoring tool that collects data on actions performed by users. Models (called profiles) are then generated and exploited, on one side, to provide feedbacks to Teachers on the validity of specific LOs and applications, and on the behaviour of Learners. On the other side, they are exploited to instruct automated tutoring agents, enabling them to suggest and guide Learners throughout the fruition process (see Di Nitto et al. 2003). The architecture in composed of three sub-modules (Fig. 5).
Teacher
Learner Instrumented Fruition Environment (IFE)
Tutoring Module (TM)
Validation Module (VM)
Row usage data and metadata
Profiling Engine (PE)
Profile DB
Figure 5: Tutoring and Validation architecture.
The Profiling Engine (PE) generates profiles starting from usage row data collected by the Instrumented Virtual Campus Fruition Environment (IFE), and LO metadata. The Validation Module (VM) provides Teachers with reports and graphics about the performance of the Virtual Campus platform and about learning behaviors of their students. The Tutoring Module (TM) provides Learners with personalized suggestions. As an example, when a choice between two or more alternative LOs is given to a Learner, the Tutoring Module can highlight the most appropriate, based on the Learner’s profile. Profiles describe Learners, LOs, and applications, respectively. The Learner’s profile is based on a “behavioral analysis”, starting from row data collected by the IFE and concerning usage. Actually, usage row data is derived from the applications Learners need to use in order to exploit LOs. The following applications, developed at Politecnico di Milano, have been considered: • PeerVerSy: a peer-to-peer versioning system (see Balzarotti et al. 2002). Learners use it for the collaborative development of digital artifacts. • WebTalk: a 3D environment (see Barbieri et al. 1999) for collaborative browsing. Learners use it to exploit LOs that need simulation of co-presence in a “class” (possibly with a Tutor acting as a guide). • Lexi.NET: a full SCORM environment. Learners use it to exploit standard SCORM LOs. • WebBoardExport: provides forum and chat facilities that, starting from the structure of a LO (typically, a Complex LO corresponding to a whole course or to a course module), automatically configures a virtual discussion environment on specific topics associated to the LO. WebBoardExport is based on WebBoard, a commercial application. The IFE also collects a subset of LO metadata stored in the Virtual Campus Repository. In particular the IFE focuses on metadata referring to time and to collaboration. Usage row data and LO metadata are aggregated and processed in order to derive the Learner's profile. In particular, the profile is composed of three categories: • Knowledge profile: this category keeps trace of assessments, time spent inside LOs (actual fruition time), failed tests, etc. • Learning style: this category keeps trace of collaboration/communication behaviors of Learners while exploiting LOs. A Learner may prefer LOs that do not imply collaboration or communication (e.g. a PDF text, PowerPoint slides, etc). On the other hand, Learners may prefer to exploit LOs in which they have to study in a collaborative way (e.g. browsing a virtual space with WebTalk or developing a digital artifact with PeerVerSy). • Behavioral style: this category contains information about the ability of Learners to correctly exploit the learning modalities and the applications the system provides. As an example, Learners that exhibit a collaborative learning style but do not use collaborative and communicative functionalities in a proper way (e.g. they do not send any chat messages while using WebTalk) are actually considered not able to exploit collaborative LOs. The LO profile aims at evaluating the effectiveness of a given LO and is computed using Learner's profiles. In particular, the profile for a given LO is calculated considering the profiles of Learners who have exploited it in the past. The application profile aims at evaluating the effectiveness of a given application and is computed using LO profiles. In particular, the profile for a given application is calculated considering the profiles of LOs that require that application in order to be exploited by Learners.
8. Conclusions and future works Virtual Campus focuses on innovative scenarios, flexibility and reusability. First of all let us consider reusability. Since all the learning resources are thought of as LOs, the metadata specification and the aggregation mechanism are independent of the learning material granularity: the definition and the aggregation of a set of lectures in a course, or a set of courses in a curriculum are performed in exactly the same way. Exploiting the customization process allowed by our three-level model, LOs can be adapted to specific Teachers’ requirements. In particular, while the Reusable Level representation allows Authors and Organizers for a high-level definition, the Didactical Level language allows Teachers for a fine customization of the permitted didactical paths. Metadata are another key aspect in Virtual Campus. We try to use as many of them as possible not only to allow for searching functionalities, but also in order to perform more sophisticated tasks. In particular, metadata are exploited by the Tutoring and Validation module. Comparing metadata description (the “expected” data) and gathered usage
data (the real data), the system is able to evaluate the effectiveness of LOs and applications. Moreover, Learners’ preferences and weaknesses can be inferred and exploited in order to give suggestions. The design of Virtual Campus has been strictly influenced by flexibility, emphasizing the importance of the integration of traditional learning scenarios and innovative opportunities. On one side, the Virtual Campus project tries to “augment” traditional scenarios with distant and virtual communication and cooperation without loosing aspects strictly related to real interactions (as an example, live classes). On the other side, the project focuses on benefits related to “machine mediation”. For example, think of the possibility to observe and exploit data regarding students’ behavior. Flexibility in Virtual Campus also means personalization and modifiability of interactions, itineraries and scenarios. Encouraging the implementation of heterogeneous pathways (live classes and electronic learning material), Virtual Campus tries to provide the best of both worlds. The system is currently still under development even though various experiments have been completed and evaluated. The research team’s future agenda includes further investigations of the interactions among the developed e-learning system and external services (especially automated testing tools). The Tutoring and Validation module will be improved, supporting Authors during the design of Complex LOs (as an example, suggesting them the “best” LO to be reused in a given context). Finally, we plan to support SCORM 1.3 in a future release of the environment.
References WfMC (1999). Workflow Management Coalition - Terminology & Glossary. Technical Document Number WFMCTC-1011. L. Mainetti, M. Monga, and L. Sbattella (2002). A Virtual Campus For Tethered And Untethered Scenarios. Proceedings of Frontiers in Education Conference (FIE), Boston, Massachusetts. S. Fischer, A. E. Saddik, and R. Steinmetz (2001). Reusable multimedia content in web-based learning systems, IEEE Computer, July. IEEE Learning Object Metadata Working Group (2002). Draft Standard for Learning Object Metadata, ltsc.ieee.org. Advanced Distributed Learning (2001). Sharable Content Object Reference Model (SCORM) Version 1.2, http://www.adlnet.org. D. Balzarotti, C. Ghezzi, and M. Monga. Freeing Cooperation From Servers Tyranny (2002). Proceedings of Workshop on Web Engineering and Peer to Peer Computing, Pisa, Italy. T. Barbieri, and P. Paolini (1999). WebTalk: a 3D collaborative Environment to Access the Web. Proceedings of Eurographics '99 Conference, Milan, Italy, September. E. Di Nitto, L. Redaelli, L. Sbattella, and R. Tedesco (2003). Tutoring and Validation in the Virtual Campus Environment. Proceedings of International Workshop on Interactive Computer-Aided Learning (ICL), Villach, Austria, September. Centra (2003). Centra web site. http://www.centra.com/. BlackBoard (2003). Blackboard web site. http://www.blackboard.com/. A. Steinacker, A. Faatz, C. Seeberg, I. Rimac, S. Hrmann, A. E. Saddik, and R. Steinmetz (2001). Medibook: Combining semantic networks with metadata for learning resources to build a web based learning system. Proceedings of the World Conference on Educationnal Multimedia, Hypermedia and Telecommunication (EDMEDIA), Tampere, Finland. J. Lin, C. Ho, W. Sadiq, and M. E. Orlowska (2001). On workflow enabled e-learning services. Proceedings of the International Conference on Advanced Learning Technologies (ICALT), Madison, Wisconsin, USA. G. de Jesus Hoyos-Rivera, R. Lima-Gomes, and J. Courtiat (2002). A flexible architecture for collaborative browsing. Proceedings of the Eleventh International Workshop on Enabling Technologies: Infrastructure for Collaborative Enterprises (WETICE), Pittsburgh, Pennsylvania, USA. Dolog, R. Gavriloaie, W. Nejdl, and J. Brase (2003). Integrating adaptive hypermedia techniques and open RDFbased environments. Proceedings of the The Twelfth International World Wide Web Conference, Budapest, Hungary.
