2009 Fourth International Conference on Internet and Web Applications and Services
Connecting learning object repositories strategies, technologies and issues
Fredrik Paulsson Dept. of Interactive Media and Learning (IML) Umeå University Umeå, Sweden e-mail: [email protected]
Abstract – The rapid growth of digital learning metadata describing them) are stored in a repository2 in resources has brought forward a number of issues order to facilitate the process of browsing DLR. While concerning availability, distribution and use. Altogether, repositories can be a good approach to indexing and the issues are a mix of interplaited technological and cataloging digital learning resources, they also introduces a pedagogical considerations. Some of those issues, mainly couple of new problems that need to be addressed: related to repositories and the distribution of digital learning resources, are examined and described in this • Descriptions of digital learning resources need to be article. A particular focus is put on how resources can be made according to a common standard and/or described and indexed using metadata, and on how Application Profile [1] in order to be useful in a access to digital learning resources can be improved and broader context. facilitated through federation and/or harvesting of • Repositories tend to become isolated islands that are metadata in order to tie several repositories together to often not indexed by search engines such as Google. provide a service that offers one single access point. The This means that users need to perform searches in study also examined how this single point of access can several different places in order to cover several be moved closer to the user (i.e. to the environment repositories, meaning that they need to be aware of where digital learning resources are used) through where to find the different repositories. simple federation of the service, enabling access to the network of repositories from any virtual learning • Repositories are often separated from the rest of the environment. The study was carried out by experiments virtual learning environment (VLE) instead of connected to a real-life case. The study concludes in becoming a transparently integrated part of the several suggestions for how access to digital learning VLE. This problem is mainly related to how VLEs resources can be enhanced, as well as in the are designed and implemented (Paulson, 2008) and identifications of a couple of new issues that need to be it needs to be solved. addressed by future research. Keywords: Learning Object Repository; OAI; Technology The Spider3 project touches upon the first problem, and Enhanced Learning; federation, LOM, RDF address the second and the third problem in more detail and from a general perspective. Hence, the methodologies 1. INTRODUCTION applied by the Spider should be applicable in other, similar cases, as well as being applicable to repositories for digital The amount of digital learning resources (DLR)1 are resource in general. increasing rapidly and a huge amount of commercial, as The remainder of this paper will review some related well as open educational resources (OER), have been made work and describe how the above problems were addressed available. However, this rapid growth of digital learning and how the solutions were designed and implemented in resources also means that it becomes harder and harder to order to develop a broker service for repositories called the get an overview and to search and find suitable digital Spider. learning resources. At the best, digital learning resources (or 2 Repositories for digital learning resources are often referred to as 1 Digital learning resources are sometimes referred to as Learning Objects. Learning Object Repositories (LOR), a term that will be used in this paper Learning Objects are regarded as a subtype of digital learning resources, henceforth. with specific properties, in the context of this paper, as discussed by 3 The official Swedish name of the project is “Spindeln”, which translates Paulsson (Paulsson, 2008). to the Spider in English.
978-0-7695-3613-2/09 $25.00 © 2009 IEEE 583 DOI 10.1109/ICIW.2009.109 The next section of the paper briefly describes the own infrastructure, i.e. their local portal, LMS, intranet or objectives, intentions and deliverables of the Spider project, whatever it may consist of. followed by a short review of some previous and related work. The review is followed by a section that describes 2. METHOD how the problems described in the introduction were The Spider project is a real case and the results presented addressed, and how the resulting solutions were designed in this paper have been applied to a Swedish national DLR and implemented. The paper ends with a brief discussion of repository broker that now runs as a service for Swedish the results and the contributions made by the research schools and offered by the Swedish schoolnet since april presented in this paper, followed by a couple of suggestions 2008. As this paper describes a real case study, most of the for future works that address some of the new issues empirical data, that constitutes the basis for the paper, was encountered during the project. collected in experimental studies, mainly using rapid prototyping. Prototype implementations have then been 1.2 The Spider: background and objectives consolidated and stabilized in order to reach production The main objective of the Spider project was to offer a strength and the fact that the Spider is now implemented as single point of entry for searching DLR in repositories a national service can be regarded as a proof of concept for targeting Swedish schools. Existing digital repositories the chosen approaches. needed to be interconnected in a “repository network” in order to accomplish the infrastructure needed for such 3. RELATED WORK service. In order to connect repositories in a network, a There is a significant amount of research that is, in one couple of basic conditions had to be met. Firstly, there is a way or another, related to the research presented in this need for a common way of describing digital learning paper. The reason is not that the same things have been done resources. Hence, common vocabularies and common several times before, but instead that the work underlying metadata standards and (preferably) a common Application the spider is based on related research from several fields. Profile [1] are needed to ensure that resources are described This work has been combined and applied in order to in a uniform way and that concepts and terms are used in a address the issues described in this paper to solutions to new consistent and coherent way. Secondly there is a need for a problems, and by that also produce an added value. common way of connecting repositories in terms of Examples of related work is found within the metadata field, communication interfaces, communication protocols, data such as the work done by Nilsson et al [4], describing how formats and query languages. These need to be simple and the Learning Object Metadata (LOM) standard [5] can be generic enough to be implemented by all targeted represented using RDF, which is the data format used repositories, but still powerful enough to support and make natively by the spider. However, Nilsson et al. [6] also good use of the chosen Application Profile. points out some serious problems with this binding One of the reflections that can be made in relation to the approach, which has recently resulted in a draft for a issues described above is that even though the solution to modified approach [7]. Other related work has been done accomplish a network of repositories is a technical by Duval et al. [8], in sorting out some basic metadata implementation, a large part of the problem is related to issues for DLR that are of relevance for the research consensus building around protocols and metadata. In part, presented in this paper, as well as the research on (and these issues are solved by the use of standards, but there is definitions of) Application Profiles [1]. Application profiles still room for interpretations that may lead to design were used to adapt the chosen metadata standard (in this decisions that make different implementations differ in ways case LOM) to the specific conditions for use by Swedish that may cause interoperability problems. However, this schools. This adaptation was accomplished by the paper mainly focuses on the technical solutions (assuming development of the SWE-LOM Application Profile that consensus was reached already) and how technology is (Paulsson, unpublished). The spider also makes use of the able to fulfill the pedagogical needs of the end users - i.e. LRE/Melt Application Profile in order to comply to the the needs of pedagogical practitioners. A reason for such requirements of the Melt project, described by Vuorikari & focus is the crucial, and often overlooked, connection van Assche [9], to which the Spider was connected as a between the technical approach and implementation and the Swedish node. offered pedagogical possibilities [2]. As mentioned above, there is a justifiable amount of The original intent with the Spider project was to only related work on metadata and repositories and a much of make the service available through a portal offering the this is found within the field of library science, such as in search service, but previous experiences show that it is hard [10, 11 and 12]. However, in recent years technologies and and resource consuming to make users (i.e. teachers and experience from library science have been applied to students) find and use a portal in order to search for digital Learning Object Repositories and the organization and learning resources [3]. It was therefore decided that the management of digital learning resources, see e.g. [13, 14, service should also be offered by a “widget” that makes it 15, 16, 17 and 18]. possible for schools to integrate the Spider service into their
584 Other work of relevance studies issues and approaches to 4.1 The Spider service design connecting repositories in repository networks, such as work As previously mentioned, the basis for the Spider service related to the Open Access Initiative (OAI4). Especially is the SCAM repository framework, which it mainly uses relevant is work that focuses on metadata harvesting using for storing metadata harvested from archives in the Spider OAI-PMH protocol, such as by Van de Sompel [19], Warner network, using the OAI-PMH protocol. However, SCAM is [20] and Halbert et al [21]. Relevant for this paper is also also set up to serve as a hub for a repository federation. The the work in the Federation of Internet Resources for federation is currently using the Fire framework. In practice Education (FIRE5) project where an approach to federated Fire is an implementation of SQI using the Java Message searches in LORs is suggested based on the Simple Query Service (JMS) [22] [23]. A “Fire plug-in” was developed for Interface (SQI) [22] [23]. Both OAI and FIRE are used by SCAM, in order to support federated Fire queries. This the Spider. means that each query that is sent to the Spider triggers two different actions: one is a search query to the harvested 4. SETTING UP THE SPIDER metadata that is stored in the SCAM repository and the The Spider brokerage service was set up mainly by using other action is a a federated query to the repository several existing technologies and software platforms. The federation using the Fire model for federation (see figure Standardized Contextualized Access to Metadata (SCAM) 1). system was used as the basis for the service [24]. In short, SCAM can be described as a framework for setting up advanced LORs. SCAM is basically a framework for metadata storage and management for repositories. In order for SCAM to work as a metadata broker, support for the OAI harvesting protocol needed Fire protocols were added. The main difference of SCAM and most other LOR platforms is that SCAM uses Semantic Web technology (i.e. the Resource Description Framework, RDF) for the representation of data, as described by Decker at al. [25]. The main reason for using the Semantic Web approach was to create the conditions needed for advanced metadata Figure 1. A schematic view of the function of the Spider service. management where machine processable semantics can be utilized to create sophisticated metadata driven applications The results of these two queries are then compiled and that makes use of the ontological relations represented by presented to the user as a search result, i.e. a list of DLRs the RDF data. One example of this is the Annofolio, where from different repositories, matching the query. digital learning resources were connected to the school The Spider repository federation is currently using the curriculum by the use of Semantic Web annotations [26]. SQI and the LRE Query Language, LRE-QL [28]. The RDF is also an excellent format for data exchange as it adds reason for those choices was that the development of the the potential of semantic interoperability to the syntactic Spider service was partly done in connection to the interoperability provided by XML [25]. In spite of this, the european Melt6 project where those technologies were Spider uses the LOM based LRE (Melt) Application Profile prescribed as a part of the reached consensus. Even though and LRE thesaurus, as well as the Swedish LOM there may be more generic approaches and more powerful Application Profile, SWE-LOM. The principles underlying technologies, this does not affect the principle of repository SCAM was described in more detail by Paulsson [24] and networks and the spider design as a model for Palmer et al. [27]. implementation, which means that this choice is only The Spider service builds to a large extent on the relevant to a small part concerning the implementation research and technologies described above. However, the details and does not affect the overall results presented in Spider service adds to this work by also exploring how this paper - in more than serving as one example of several those technologies can be used in new contexts. The Spider query languages that can potentially be used in repository also adds a couple of new technologies that were needed in federations. order to make better use of the service and to meet 4.2 The Spider Widget requirements from teachers and students, such as the Spider widget that is used to federate the service to other web sites, One of the most important issues that was addressed was intranets or learning platforms. how users should be able to interact with the service in the most suitable way from the users own point of view. Previous experience, such as from the Swedish schoolnet [3], has shown that it is very hard, as well as inefficient, to 4 http://www.openarchives.org/ 5 http://fire.eun.org/ 6 http://melt.eun.org/
585 reach teachers via web portals, which they are expected to visit in order to use a service. A combined approach was 5. CONCLUSIONS AND DISCUSSION chosen for this reason - combined in so far as that the Spider As described above, the Spider services uses the LOM- service was exposed via the Swedish schoolnet portal, but based Application Profile7 for Melt and the LRE thesaurus8 the same service was also exposed via a widget (a simple for controlled vocabularies. This choice was partly made by JavaScript API, called the SCAM JavaScript Syndication), considering synergies with the Melt project. However, the which was implemented as a plugin module to SCAM. The Spider will also support the Dublin Core Education (DC- basic idea of the widget was to let anyone implement the ED) Application Profile [30] in the near future. The DC-ED Spider service in a simple way on any webpage, portal, profile is also more suitable for use with RDF (in order to LMS, intranet or in any other web based environment, all in make better use of the inherent potential of SCAM). DC-ED good Web 2.0 spirit. Hence, the spider was developed to be is built upon an abstract model that is not document centric a service for inclusion in mashups and is in that respect a in the way that XML based metadata models, such as LOM, combination of Web 2.0 technologies and Semantic Web tend to be and by that more RDF friendly [6] [31] [25]. technology and concepts [29]. In fact, the Spider widget is Even though the Spider supports both the Melt and the the technology that is used to implement the service at the SWE-LOM Application Profiles, the metadata reality in the official Spider portal at the Swedish schoolnet. A side-effect connected repositories is in general very different. The use created by this approach is that all updates to the service of metadata in Swedish LORs is poor and most repositories will get immediate impact everywhere without any doesn't implement a metadata standard for digital learning additional work on the local level. The client side script for resources. In fact, in many repositories the concept of the JavaScript based widget is listed in listing 1. The only metadata and its related principles are more or less requirements for using the widget is to cut and paste the unknown. For this reason, a subset of metadata fields was code to the
of the webpage and an additional six chosen to serve as a “smallest common denominator” and lines of code where the search form and search results shall the smallest part of the Application Profiles that must be be exposed. implemented by the repositories in order to be connected to the Spider service. This core set of metadata (and the rest of repository federation, where the queries are sent to the some additional efforts, which not all archive owners are used is metadata mapping, i.e. the existing metadata fields corresponding fields in the Application Profile. This works to some extent, but it creates unwanted overhead and there … are cases when there is no corresponding field in the original repository and this causes the mapping to fail and7http://info.melt- project.eu/ww/en/pub/melt_project/melt_content_audit.htm#AP 8http://insight.eun.org/ww/en/pub/insight/interoperability/learning_resourc e_exchange/metadata.htm
586 the accuracy of the responses to search queries are low, services like the Annofolio and the annotated curriculum, as users will soon loose faith in the service and ultimately stop described by Paulsson [26]. using it. However, in order to gain the full advantage of a One conclusion that can be made, in relation to the first SPARQL based approach, the metadata in the connected problem that was identified in the introduction, is that it is repositories and in the Spider need to be represented using important that standards for metadata and common RDF. This is not the reality right now but might very well be Application Profiles, vocabularies, taxonomies etc. are used so in the future. Alternatively, a part of the metadata (some and that metadata quality is secured - preferable on the layers) could be represented using RDF. This would not be a archive level - in order to obtain good service quality. In perfect solution, but it could at least provide some order to achieve quality it must be made easy to index possibilities for utilizing machine processable metadata resources and it is therefore an urgent need for better and semantics. more flexible metadata tools in the sense that they are The combination of federation and harvesting for integrated with the tools that teachers and content producers implementing a repository network was proven to be rather use for production on a daily basis. Metadata may also be successful by the Spider project. The work presented in this managed at different levels where there are formal, but also paper shows that the best of both worlds can be combined in less formal, “layers” of metadata where metadata can be order to make it suitable for different kinds of archives, added by other users than trained indexers, for example by providing different alternatives for connecting the “isolated teachers and students [32]. archive islands” that were described in the introduction. An alternative (or rather a complementary) approach to Another interesting (and possibly complementary) approach manual indexing is automatic indexing. Automatic indexing would be to explore the work done within the may work fairly well for some types of resources [33] [34]. Search/Retrieval via URL (SRU) initiative [37] and However, machine indexed metadata still needs a high Sanderson et al [38] make a comparison of the SRU/W and degree of manual supervision. Alternatively, when turned the OAI protocols, and points out several synergies that are around automatic, metadata generation methods can be used relevant to the future of the Spider service in order to for automatization of the evaluation of metadata quality as achieve a better combination of harvesting and federation. It proposed by Ochoa & Duval [35]. Automatic indexing and has become quite clear during the project that complexity automatic evaluation of metadata are areas that are strong should be avoided as far as possible. For each archive that is candidates for further exploration in future research. added (especially to the federation) the complexity increases The use of a LOM-based Application Profile requires a and errors within one repository may multiply and spread to supporting query language. Even though the choice of SQI the brokerage service. For this reason, the brokerage service makes the Spider rather flexible (it does not prescribe is more likely to succeed when starting out with only a few neither the data format to be used, nor the query language) simple standard- and consensus based rules for data formats, [28] the choice of query language can potentially be a communication protocols and metadata, based on solid and limiting factor. Previously to using the LRE-QL, the Simple reliable technologies, rather than trying to solve everything School Query Language (S2QL) was used and the result from the beginning using complex and sophisticated sets were expressed using strict LOM. The use of S2QL and technology. It is better to develop the service in an its simple semantic abilities limited the search quires to a evolutionary way over a longer period in order to avoid smaller set of metadata fields (keywords, age range, and problems that are related to complexity. This is one of the language), which in turn affected the search ability, and the experiences contributed by the Spider project. potential for filtering of relevant resources, in a negative An important contribution of the work presented in this way, i.e. providing a less fine-meshed and flexible service. paper is the widget approach that makes it possible to Even though the LRE-QL is a more complete choice of a combine the portal approach with a local access to the query language, it is still limited to the context of querying service. The Spider widget is frequently used and most LOM-based metadata. This becomes a limiting factor since common is to implement it at the municipality level or at the SCAM repository and the Spider service natively regional media centers. This means that it will reach the supports RDF and have the potential to support more users in their own learning environments instead of relying sophisticated query languages that make use of the semantic on users visiting a portal. Hence, providing a way of abilities of RDF. An alternative approach would be to use a integrating access to external repositories into the VLE. This query language that is developed for querying RDF data, approach is likely to have increased the use of the Spider such as the Simple Protocol and RDF Query Language service. (SPARQL) [36]. The use of SPARQL would enhance the power and the flexibility of the service since SPARQL enables direct queries of RDF graphs and enables the response to be a RDF graphs for further processing of the search result. These are qualities that for example could be useful for curriculum mapping by connecting the Spider to
587 5.2 Future work REFERENCES The next steps for the Spider will focus on enhancing the [1] Heery, R. & Patel, M. (2000). Application profiles: mixing and metadata quality. This is a complex task since it requires matching metadata schemas - introduce the 'application profile' as a type of metadata schema. Arriadne, 25, 1361-3200. enhancements at the repository level, which is technically [2] Paulsson, F. (2008). Modularization of the Learning Architecture: out of our control. However, enhancing the metadata quality Supporting Learning Theories by Learning Technologies. Doctoral is important in order for the service to survive in the long Thesis, Royal Institute for Technology (KTH), Stockholm. run. This is especially important since one of the advantages [3] Paulsson, F. (2002). Interview with Christina Szekely, project of a structured- and metadata based approach together with manager at the Swedish Schoolnet. a brokerage service is the ability to add additional services [4] Nilsson, M., Palmér, M., & Brase, J. (2003). The LOM RDF Binding that builds upon the brokerage service, such as with the - Principles and Implementation. Proceedings of the Third Annual ARIADNE conference. Annofolio [32]. [5] IEEE. (2002). Final 1484.12.1-2002 LOM Draft Standard. Other enhancements that have priority are related to the http://ltsc.ieee.org/wg12/files/LOM_1484_12_1_v1_Final_Draft.pdf widget, in terms of extending its functionality, as well as by [6] Nilsson, M., Johnston, P., Naeve, A., & Powell, A. (2006). Towards providing more advanced alternatives. Alternative an Interoperability Framework for Metadata Standards. Paper approaches to the JavaScript-based widget was discussed presented at the the International Conference on Dublin Core and Metadata Applications, Manzanillo, Colima, Mexico. from start, but since ease of use was the number one [7] IEEE. (2008). Draft Recommended Practice for Expressing IEEE priority for the Spider widget, the most simple solution Learning Object Metadata Instances Using the Dublin Core Abstract found (even though it is not the most powerful) was chosen. Model. P1484.12.4™/D1. Future solutions that utilizes the RESTful architecture [8] Duval, E., Hodgins, W., Sutton, S., & Weibel, S. L. (2002). Metadata proposed by Fielding [39] in the upcoming version (4.0) of Principles and Practicalities. D-Lib Magazine, 8(4). the SCAM framework is planned. Another approach that [9] Vuorikari, R. & Van Assche, F. (2007). Collaborative content will be explored is a Portlet based approach that builds upon enrichment in multilingual Europe, European Schoolnet approach on educational resources. Paper presented in the proceedings of the 3rd the RESTful widget. This is to make the portal version of International Conference on Automated Production of Cross Media the Widget a bit more standardized, as well as more flexible. Content for Multi- Channel Distribution. It is important in general that the service provide alternative [10] Baldonado, M., Chen-Chuan K, C., Gravano, L., & Paepcke, A. API:s that is more advanced and built on a more solid (1997). The Stanford Digital Library metadata architecture. International Journal on Digital Libraries, 1(2), 108-121. technology base in order provide access to all existing and [11] Godby, C. J., Young, J. A., & Childress, E. (2004). A Repository of new functions that is/will be provided in the future. Metadata Crosswalks. D-Lib Magazine, 10(12), Online. In parallel, there is a need to further explore alternative [12] Robertson, R. J. (2005). Metadata quality: implications for library and approaches to federation and alternative query languages information science professionals. Library Review, 54(5), 295-300. that can help make the service more powerful and flexible in [13] Reichle, R., Wagner, M., Ullah Khan, M., Geihs, K., Valla, M., Fra, line with the previous discussion. The SRU/W and OAI C. et al. (2008). A Context Query Language for Pervasive Computing Environments. Paper presented at the 2008 Sixth Annual IEEE combination, for example, would most likely provide a International Conference on Pervasive Computing and more generic approach than the approach currently used by Communications, Washington, DC, USA. the Spider. [14] Richards, G., McGreal, R., & Friesen, N. (2001). Learning Object It is plausible that we will see other services that target Repository Technologies for TeleLearning: The Evolution of POOL schools, which build upon similar principles, in the future. and CanCore. Paper presented at the Proceedings of the Informing Science+ IT Education Conference, Cork, Ireland. This will be necessary in order to become an integrated part [15] Paulsson, F. (2005). Achieving Learning Object Repository of their learning infrastructure [2]. The Spider service is interoperability: Position paper for the LIFE workshop on Learning provided as a module (widget) for inclusion in schools Object Repository Interoperability. Paper presented at the LIFE learning infrastructure. However, schools will be able to [16] McClelland, M. (2003). Metadata standards for educational utilize the full potential of services such as the Spider first resources. IEEE Computer, 36(11), 107-109. when their infrastructures becomes truly modular. This [17] Najjar, J., Ternier, S., & Duval, E. (2003). THE ACTUAL USE OF METADATA IN ARIADNE: AN EMPIRICAL ANALYSIS. Paper would make it possible to pick and chose from a pool (or presented at the ARIADNE 3rd International Conference. toolbox) of services in order to “mashup” the virtual [18] Neven, F. & Duval, E. (2002). Reusable learning objects: a survey of learning environment that is needed for a specific LOM-based repositories. Paper presented at the the tenth ACM pedagogical situation – quite opposite to the silo-like international conference on Multimedia. learning platforms (LMSes) that are commonly in use today [19] Van de Sompel, H., Nelson, M. L., Lagoze, C., & Warner, S. (2004). [2] [40]. Resource Harvesting within the OAI-PMH Framework. D-Lib Magazine, 10(12). ACKNOWLEDGMENT [20] Warner, S. (2001). Exposing and harvesting metadata using the OAI metadata harvesting protocol: A tutorial. High Energy Physics The research presented in this paper was financed by the Libraries Webzine, 4, 1-13. European Union (via the eContent Plus) and the Swedish [21] Halbert, M., Kaczmarek, J., & Hagedorn, K. (2004). Findings from Agency for School Improvement (MSU). the Mellon Metadata Harvesting Initiative. In G. Goos, J. Hartmanis, & J. van Leeuwen (Eds.), Research and Advanced Technology for
588 Digital Libraries (Vol. 2769, pp. 58-69). Heidelberg: Springer [32] Paulsson, F. & Engman, J. (2007). Treating metadata as annotations: Berlin / Heidelberg. separating the content mark-up from the content. International [22] Colin, J.-N. & Massart, D. (2006). LIMBS: Open Source, Open Journal of Emerging Technologies in Learning (iJET), 1(3), 1-7. Standards, and Open Content To Foster Learning Resource workshop on Learning Object Repository Interoperability, Exchanges. Paper presented at the Sixth IEEE International Manchester, UK. Conference on Advanced Learning Technologies (ICALT'06). [33] Olivié, H., Cardinaels, K., & Duval, E. (2002). Issues in Automatic [23] Massart, D. How To Light FIRE. http://fire.eun.org Learning Object Indexation. Paper presented at the World Conference on Educational Multimedia, Hypermedia and Telecommunications [24] Paulsson, F. (2003). Standardized Content Archive Management – (EDMEDIA) 2002, Denver, Colorado, USA. SCAM. IEEE Learning Technology newsletter, 5(1), 40-42. [34] Cardinaels, K., Meire, M., & Duval, E. (2005). Automating metadata [25] Decker, S., Melnik, S., van Harmelen, F., Fensel, D., Klein, M., generation: the simple indexing interface. Paper presented at the 14th Broekstra, J. et al. (2000). The Semantic Web: the roles of XML and international conference on WWW, Chiba, Japan. RDF. Internet Computing, IEEE, 4(5), 63-73. [35] Ochoa, X. & Duval, E. (2006). Towards Automatic Evaluation of [26] Paulsson, F. & Engman, J. (2005). Marking the National Curriculum - Metadata Quality in Digital Repositories. Lecture Notes in Computer a new model for semantic mark-up. In P. Cunningham & M. Science, Volume 4231/2006, 372-381. Cunningham (Eds.), Innovation and the Knowledge Economy: Issues, Applications and Case Studies (Vol. 2, pp. 1731-1738). Amsterdam: [36] Prud'hommeaux, E. & Seaborne, A. (2008). SPARQL Query IOS Press Amsterdam. Language for RDF. W3C Recommendation Retrieved 10 October, 2008, from http://www.w3.org/TR/2008/REC-rdf-sparql-query- [27] Palmér, M., Naeve, A., & Paulsson, F. (2004). The SCAM 20080115/ Framework: Helping Semantic Web Applications to Store and Access Metadata. Paper presented at the the European Semantic Web [37] McCallum, S. H. (2006). A Look at New Information Retrieval Symposium 2004, Heraclion Greece. Protocols: SRU, OpenSearch/A9, CQL, and XQuery. Paper presented at the World Library and Information Congress: 72nd IFLA General [28] Massart, D. (2006). LRE-QL: Introducing an LRE Context Set For Conference and Council, Seoul, Korea. The Federated Discovery of Learning Resources. Retrieved 10 October, 2008, from http://fire.eun.org/QueryingTheLRE.pdf [38] Sanderson, R., Young, J. A., & Le Van, R. (2005). SRW/U with OAI Expected and Unexpected Synergies. D-Lib Magazine, February [29] Ankolekar, A., Krötzsch, M., Tran, T., & Vrandečić, D. (2007). The 2005. two cultures: Mashing up Web 2.0 and the Semantic Web. Journal of Web Semantics: Science, Services and Agents on the World Wide [39] Fielding, R. T. (2000). Architectural Styles and the Design of Web, 6(1), 70-75. Network-based Software Architectures. University of California, Irvine. [30] Sutton, S & Hillmann, D (2008). DCMI Education Application Profile (WD 3). Retrieved October 3, 2008, from [40] Downes, S. (2002, 11-13 November). Design and Reusabilityof http://projects.ischool.washington.edu/sasutton/dcmi/DC-EdAP-7-18- Learning Objectsin an Academic Context: A New Economy of 06.html Education. Paper presented at the eLearning una sfida per l'universita, Milan, Italy. [31] Powell, A., Nilsson, M., Naeve, A., Johnston, P., & Baker, T. (2007). DCMI Abstract Model. Retrieved September 10, 2008, from http://dublincore.org/documents/abstract-model/
589