3.2 Design and Development of Software System for Resource Service Platform
In recent years, various educational institutions in China have built a considerable number of educational resource repositories. However, owing to great differences in development standards, management specifications, structural framework, adopted hardware and software platforms, etc., resource repositories are mutually independent and lack uniform standards and interfaces. There is no interoperability between application programs, which make intercommunication and resource sharing difficult between systems. To solve these problems, design and development of the platform software system must include a thorough study of the application goals and client demands of the platform, a sound overall framework for the platform and subsystem functions, and specialized research on key related technologies.
3.2.1 Design and Development of the Platform
Based on the resource center's integration and application features, we structured the system as a two-level platform, that is, the master center level and sub-center level. The system offers uniform management of directory information and distributed storage of resource entities, provides users with customized service, and facilitates data exchange between the platform's two levels including user information, directory information, supply and demand information, transaction information, resource recommendation, interactive communication, resource promotion, etc.
This system design accommodates the demand for sub-centers to manage their own resources autonomously while facilitating connectivity by implementing identical standards and interfaces for the master center and sub-centers. The design is distributed, standardized, interoperable, and offers transparent user access. (Figure 3).
Figure 3 Framework of Network Education E-learning Resource Center Platform
The platform consists of such subsystems as a resource repository system, learning platform, transaction system, navigation system, credit bank, individual development files and a human resources marketplace. It also contains tools such as on/offline cataloguing tools, import and export tools, directory harvesting tools, network course authoring tools, statistics tools, certification management of sub-centers and external application interfaces (Figure 4).
Figure 4 Subsystem Constitution and Relevant Tools of E-learning Resource Center Platform
3.2.2 Research & Development and Application of Key Technologies
In the process of developing the platform, we studied and applied the following key technologies: storage and management of massive resource entities, WCF distributed data communication, metadata interoperability, quick search of massive information resources, self-adaption of compound documents, self-adaption of user terminals, self-adaption of layered media, etc.
(ⅰ) Massive Resource Storage
Based on the operating systems of custom high-end servers, large-scale scalable DFS and database technologies have been adopted to achieve the unified storage management of shared resources from a physical perspective and support the continuous online extension of storage capacity. The capacity for 5000 courses and 50 TB of storage is ensured.
(ⅱ) Metadata Standards Application and Heterogeneous Resource Conversion
A structured review of resource metadata specifications has been conducted, and a series of tools was used to edit, convert, import and export metadata, making it easier to store and manage resources. Metadata standards were applied.
(ⅲ) Resource Synchronization
Technologies including distributed computing, utility computing, virtualization and load balancing have been adopted to achieve synchronous and unified management of widely dispersed e-learning resources.
(ⅳ) Resource Retrieval
Natural language processing technologies such as Chinese intelligent segmentation and semantic analysis have been used to optimize retrieval speed to the milliseconds level and achieve an accuracy level of 98%, ensuring quick and accurate access to resources.
(ⅴ) Transparent User Access
Technologies such as Remote Procedure Call (RPC) and Representational State Transfer (REST), etc. have been applied to build a unified authentication platform (user center), achieve single sign-on for every sub-center and provide resource sharing services.
(ⅵ) Large-Scale Concurrent Users
Access performance has been improved to support large-scale concurrent access using the following three methods: dispatching access according to a routing mapping table, multi-server clustering and load balancing, and providing dedicated servers to serve heavily accessed resources.
(ⅶ) Building Interfaces to the Online Learning Platform
Interfaces related to the mainstream learning platform have been established to connect the resource platform with the main learning platform.
(ⅷ) Multiple Terminal Access to Resources
The system supports multiple-terminal access and implements a web content analysis tool and webpage self-adaption strategy. While preserving the rich content and intelligibility of the webpage, the system minimizes the burden caused by users' access and implements a self-adaptive resource transmission system that can support multiple terminals. It is primarily designed for PC access and also supports the access of digital satellite televisions and handheld terminals.
(ⅸ) Network Information Security of the Resource Sharing Platform
In order to ensure the safe operation of the e-learning resource center, we launched a special research project entitled Research and Development of Network Security and Information Management Systems. Grounded in the actual conditions of the project and based on application needs, it analyzes important components of sub-center node repositories and develops a repository network protection system, application security management system and protection system for confidential information. The project includes security management, static web protection, cross-site attack prevention, SQL injection filter control, etc.
In addition, to optimize resource allocation, promote high-quality resource sharing between schools and provide a free, open and lifelong learning environment to students, the project included research for a credit bank system. The credit bank system would bridge degree and non-degree education, accumulate and transfer credits of different education forms, and incorporate main services such as credit certification, credit transfer, credit accumulation, credit retrieval, etc.
3.3 Integration and Evaluation of Massive Resources
3.3.1 Integration of Massive Resources
(ⅰ) Establishing Uniform Standards of Resource Classification and Metadata
The e-learning resource center serves as a public service platform, and its ability to serve diversified users is closely tied to the sources, types, and scale of the resources added to the resource center system. Resource standardization is thus a precondition of e-learning resource sharing and reuse.
Currently, different educational institutions generally follow different technical standards in constructing resource repositories. They use different metadata standards for description and different classifications for storage and management, which inevitably creates difficulties in integrating and sharing resources.
In the process of project implementation, we researched the requirements for resource classification and main classification methods currently used in the field of educational resource sharing. Factoring in the needs of the project, we confirmed the basic principles of resource classification and designed the resource classification system for the resource center. This includes course classification and learning resource classification.
Degree course classification was formulated based on the MoE's catalogue of undergraduate, vocational and secondary disciplines, and GB/T13745-2009 (Classification and code of disciplines). It includes 13 first-level classifications and nearly 200 second-level classifications. Non-degree course classification was based on application research and includes 18 first-level classifications and 108 second-level classifications.
Learning resource classification was formulated based on the Chinese E-Learning Technology Standard (CELTS-41.1), covering media material classification and media teaching attribute classification. It includes 9 first-level classifications and 69 second-level classifications.
As for resource metadata standards, the project team researched domestic and international educational resource metadata standards (Dublin Core, LTSC LOM, ADL/SCORM, etc.) in light of China's actual application needs, and established a resource metadata standard application scheme (defining 44 metadata attributes on the basis of GB/T21365-2008). An application for standards conformance testing has already been submitted to CELTSC.
(ⅱ) Forming a Complete Set of Resource Integration Specifications and Procedures
- Formulating Specifications for Processing and Cataloguing Resource Inventory
These include standardizations in the aspects of resource processing and organization, metadata scheme, resource identifiers, description formats, structure, etc.
After fully considering user demands, the master center defined the following four general specification categories: media material resource cataloguing specifications, course resource cataloguing and correlation specifications, video resource clipping and processing specifications, and media material resource processing specifications. On top of these general specifications, nine special specifications integrating the individual characteristics of the resources were developed, including a "national excellent courses (including network excellent courses) resource cataloguing specification", "media material resource cataloguing specification for open courses of the OUC", etc. Resource editors catalogued resources according to uniform specifications and assured the quality of resource inventory.
- Establishing Resource Integration Specifications and Procedure
Based on practical exploration, the project team established the resource integration procedure from resource demand research to plan, import, process, catalogue, check, release and application. For the demand research phase, questionnaires, interviews, etc. were used to probe the resource demands of sub-centers, institutional users and individual users, and thus determine the objects and scope of resource integration. Based on these demands, resources could be collected through various ways such as obtaining for free, transaction and exchange, and cooperative construction. The resources could then be added to the repository according to processing and cataloguing specifications. These resources will provide resource support both to teachers and students (Figure 5).
Figure 5 Procedure of Resource Integration of the Resource Center
(ⅲ) Formulating a Business Mode for Resource Integration and Sharing Services
According to the characteristics of different e-learning resources, we designed three business modes of resource integration: resource entity integration, resource catalogue integration and heterogeneous repository resource integration. Resource entity integration refers to the cataloguing and repositing of digital resource entities. Resource catalogue integration refers to the cataloguing and repositing of resource attribute information and link addresses. Heterogeneous repository resource integration refers to working with heterogeneous repositories to integrate resources according to metadata mapping and the like.
The integrated resources will be provided to users in the following three forms: courses, media materials and topical resources. Course resources include "course resource packs", split and restructured courses, and URL courses. Media materials refer to resources that have been disaggregated, reducing resource granularity and facilitating partial reuse. Topical resources are media or course resources reorganized around certain topics, which are easier to navigate and use by category.
So far, under the guidance of the above specifications and procedures, we have integrated the course resources of 3600 courses from radio and TV universities, traditional colleges and universities, network colleges, vocational schools, training organizations, publishing organizations, and Hong Kong, Taiwan and foreign educational institutions. The total capacity of the resources exceeds 16TB. Over 45,000 material resource items in the form of video, text, courseware, test/exercise, animation, cases, etc. have been disaggregated, processed, catalogued, and reposited. The content of these resources covers degree education, rural education, community education, etc. As the pilot application of the project progresses, the total amount of resources will be increased substantially.
3.3.2 Evaluation and Certification of High Quality Resources
E-learning resources are diverse in form, vast in quantity, rich in content and are increasing rapidly. However, they vary greatly as to quality and effective lifespan, and are updated frequently. One of the important problems of resource repository construction lies in how to maintain an effective “metabolism” for the resource repository system, promoting the high quality resources and eliminating the inferior resources.
China currently lacks an authoritative, dedicated resource appraisal organization and mature evaluation standard system. Under these circumstances, the project set out to build an e-learning resource evaluation system from the perspective of learners. The details are as follows:
(ⅰ) Study of Learners and Online Learning
The learner-oriented evaluation system firstly studies learners and their online learning behavior to obtain important information such as models of learners' personal characteristics, online learning habits, etc. Learners are then divided into several abstract categories.
(ⅱ) Study of Network Education Resources
Based on the Network Education E-learning Resource Center Construction project's prior research conclusions, the types, forms, content features, etc. of network education resources were analyzed to find the characteristics of resources that are suitable for learners’ autonomic learning, and test resource repositories were built accordingly.
(ⅲ) Design of Network Education Resource Evaluation System that Supports Personalized Learning
The network education resource evaluation system was designed based on learners and their online learning habits (complete index system includes approximately 6 dimensions and 46 second-level indices), and has been evaluated preliminarily with such empirical methods as the expert method (teaching design experts, teaching content experts and educational technology experts).
(ⅳ) Algorithm Study
An intelligent algorithm for comprehensive evaluation of e-learning resource quality was designed from the learner perspective. The algorithm ignores characteristics such as categories, heterogeneous distribution and differing standards and lets learners actively or passively evaluate the quality of e-learning resources from the dimensions of resource usage frequency, passive evaluation and active feedback of learning effectiveness, cost efficiency, etc. Extracting this feedback data, we obtain an evaluation of the quality of e-learning resources from different learners.
(ⅴ) Study of Plug-ins in Network Education Resource Evaluation
On the basis of research performed for the network education resource evaluation system, advanced J2EE technology was used in developing relevant plug-ins for the evaluation system. Finally, the network education resource evaluation tool was initialized and applied to a real environment to measure its self-adaption ability. The tool will be optimized accordingly for widespread application.
4．Operating Mechanism and Pilot Application of E-learning Resource Center
In order to build a successful e-learning resource center, it is critical to put in place a long-term operating mechanism, that is, ensure the sustainable growth of resource volume and the sustainable development of resource sharing services. Establishing a mutually beneficial, win-win resource sharing mechanism and "sub-center system" is necessary to achieve this goal.
The sub-center system consists of the master center and a large number of regional, professional, collegiate and training sub-centers. These centers jointly constitute the resource public service platform, the far-reaching resource catalogue center, resource transaction center and resource coalition. Meanwhile, research has been conducted to formulate a public service mode and operating mechanism for improving high quality resource sharing.
In the sub-center system, various sub-centers serve as resource providers as well as resource users. The master center and sub-centers have equal status, and together they form a resource coalition and decide the application mechanism for resource sharing. This guarantees the diversity of the resource sources and the efficiency of resource sharing (Figure 6).
During the actual operation of the resource sub-centers, all educational institutions including radio and TV universities at every level, network education colleges of regular institutions of higher education, vocational colleges, local distance education centers, training organizations, publishing organizations, learning enterprises, learning towns, learning communities and other lifelong educational institutions, may apply to be a sub-center or demonstration center. The master center would provide them with resource repository management and application software, system deployment, application technical support and other services.