In the Hong Kong Institute of Education (abbr. HKIEd), a vast amount of lesson plans for primary and secondary teaching practices are prepared by our pre-service teachers. These valuable documents are being wasted each year. One reason for this is that there are lack of formal mechanism to keep them. The other reason is that the non-standardization of these electronic documents prevents an effective exchange and retrieval of their contents. In this paper, a formal markup language for lesson plans is proposed and developed using XML. The formalization for the various components of lesson plans helps to foster a non-ambiguous data model. Using XML technology, not only we can relevantly query the lesson plans, we can display their contents in a variety of formats and styles using web-based technologies such as XSL and XSLT.

Many tangible computational construction platforms are designed with new technologies and children's learning theories to assist children in learning computational concepts. This thesis explores how children learn computational concepts through these platforms, by focusing on three categories of tangible computational platforms: tangible programming systems, computational toolkits for crafts making, and computational interfaces for story creation. Our rationale for choosing these three categories is motivated by the hands-on learning theories argued by three influential educators: Montessori, Froebel and Vygotsky. Given this motivation, we design three dimensions of case studies: abstract thinking through functional blocks construction, creativity through crafts making and expression through stories creation, for analysis of children's learning of computational concepts vertically through three main categories of computational construction kits respectively. Each dimension of empirical studies obtains qualitative and quantitative results, and the results indicate that children have positive learning experience on each category of computational construction kits in different aspects. For the study in building functionality, we use a tangible programming system to study how children understand the abstract concepts by building functional blocks to simulate functions of smart clothing and flows of story, as well as computational concepts such as looping and branching. To study creativity through crafts making, three kinds of computational craft platforms are adopted as case studies. These are the Lilypad Arduino with TeeBoard platform, i*CATch apparel platform and i*CATch robotic platform. The focus is on crafts making in robotic and apparel domains, integrating traditional materials such as paper and cloth with the electronic devices. Through these three different design approaches of the computational craft platforms, we gain a deeper understanding of how children use

The research in applying data mining to education became very intensive, and web-based education continuously proliferated. Between learners and the information content, the interaction adaptation is the focus of a lot of studies, and then the framework of personalized e-learning is introduced and used in various application domains as a generic term. Many models of e-learning personalization are developed to collect the experiences generated in the learning process and analyze the characteristics based on the combined information of learners. Experiments are conducted by using statistical methods, data mining and semantic web mining to analyze LMS data, and the results prove the effectiveness in real-world practice. The concepts of learner profile and learning objects content are introduced for better utilizing any technologies methods to do the modeling. The network and repository of learning objects are designed to center the learner and monitor all interactions. At last, personalized services are adapted with optimal operations, content and sequence according to the well- modeled profile. This paper proposes a personalized e-learning system model which integrates data and text mining techniques for educational services personalization into the popular LMS software called Moodle. A number of mature technologies and methods are used with some customizations to provide the necessary functionalities and solve the integration problem of system. The model is implemented by developing plug-in and background programs of data collection, processing, data mining algorithms, and web user interfaces. At last, the solution system with real and simulated data is tested, and an assessment of student classification is conducted.