BILDU: Compile, Unify, Wrap, and Share Digital Learning Resources

Introduction The reuse of existing learning materials is an important topic in current research (Zimmermann, Meyer, Rensing, & Steinmetz, 2007) where learning objects play an essential role. One of the problems that must be overcome for the normalization of learning objects utilization is the difficulty of finding them (Bates, 2005). Obviously, search engines can solve the problem of seeking and retrieving learning objects, but there is a lack of suitable search engines in this field. Nowadays, the use of search engines has widely contributed to the success of new learning processes, e.g. thanks to search engines users can carry out their own personal research to collect learning materials. But, despite the Internet providing an inexhaustible amount of learning materials, the shortage of learning objects within all the learning resources available on the Internet hampers the effectiveness of learning object search (Taibi, Gentile, & Seta, 2005). Concepts of learning object and learning resource should be distinguished. For the purposes of this paper, we define learning object as an independent digital resource packaged in accordance with the rules set by any of the international standards (e.g. SCORM, IMS, or LOM), while we define learning resource as any digital content that could be used for educational purposes (e.g. bibliographic references, research papers, book chapters, multimedia material, and learning objects). Figure 1 represents the percentage of each learning material type in the repositories indexed by OpenDOAR (http://www.opendoar.org). The chart shows that only 13% of all the learning resources are learning objects ("Content Types," 2007). Therefore, it can be assumed that learning objects are a small part of learning resources. As learning objects are based on learning resources, we think that the improvement in the search of learning resources is a prerequisite to facilitate the creation and reuse of learning objects (Portillo, Romo, Benito, & Casquero, 2007). However, when trying to find learning resources, current search engines have the following limitations: * Great amount of non-meaningful results. It requires some time to filter which digital resources really relate to learning resources; i.e. the number of results is not indicative of quality, e.g. a broad Google search delivers a very large number of resources of low value, whereas an Ovid search (http://www.ovid.com), while only returning one result, can point to a resource of very high quality. * Lack of specific indexes for learning resource repositories. There are a great number of information sources related to e-learning on the Internet, and it is necessary to assess the quality and reliability of the material they contain. * Need for more advanced user interfaces to launch searches and present results. The traditional text box where keywords are typed and the usual list of hits returned by a search engine should be enhanced. With respect to queries, the use of a more advanced syntax does not always help to reduce the large number of returned results. Regarding the list of results, ranking mechanisms try to sort them and show best matches first, but this notion of relevancy is typically a score computed out of elements (e.g. number of occurrences of a keyword, proximity of keywords, etc.) that do not necessarily represent user preferences (Abel, Herder, Karger, Olmedilla, & Siberski, 2007). Besides, there is not usually information about relations among different results. * Shortage of social networking services. Search engines do not facilitate collaboration between people with similar interests. This paper focuses on the whole set of learning resources, rather than on the subset of learning objects, and explores solutions to enhance access to digital resources that are likely to be used as learning resources by various communities. …