Persistent Knowledge Acquisition for Educational Purpose : An Open Distance Learning Website for German Literature and Language Scientists

Building learning packages and educational software often forces a decision which learning system/software to use, or to develop a system for particular needs. The disadvantage of many systems is, that they are „data islands“ in the rapid stream of information flow. Exchange of data especially in learning and multimedia systems was not enough considered in the past, and will be a crucial criterion in the future for preserving valuable information once collected. Extensible Markup Language (XML) is a very interesting approach to solve this problem of persistent data storage, especially when combined with metadata information and platform independent tools. We used these technologies to build a framework for an Open Distance Learning Course (ODL) „Literatur in der Wiener Moderne“ dealing with Austrian Literature. It was desired to record the data required for publication in a highly generic form, which is difficult in the case of lecture data and multimedia information. Nevertheless the aim was to separate data from presentation, for gaining flexibility for future use of the data in arbitrary other (scientific) projects. Moreover as the internet and multimedia technology is evolving with an enormous speed, our approach allows easy adoption of data to future internet or multimedia technology. 1. "Data Islands" and the Need of Data Exchange 1.1. Proprietary Systems, „Learning Software“, „Authoring Systems“ The situation of multimedia authoring systems is a very dynamic one and innumerable software vendors try to place their products in this emerging market, with the consequence that many different proprietary systems are available today. The problem is that the ability of data exchange is not always a prior function of these systems. Some producers might have in mind, that a good ability of data exchange allows the users to „easily“ migrate to a system of an opponent. The consequence is, that nearly all systems use proprietary, and even worse, usually binary (non documented) data formats. This is especially true for systems from „smaller“ vendors or systems that were designed for certain universities or training institutes. The lack of common standards makes data exchange between different systems a torture. Sometimes lots of conversion tasks have to be performed manually, if possible at all. At least many of these systems allow to structure the data in a generic way, which offers the possibility for future reuse in different contexts. Sometimes exchange is only possible in very unstructured data formats like „plain text“, HTML etc. Information loss can not be avoided in most of these cases. Figure 1 illustrates this problem: At the moment (blue islands) different systems do exist with proprietary data storage, consistent, and 1 Institute for Software Technology, Vienna University of Technology with a (hopefully) generic data structure, but data exchange or use in „Future“ Systems is usually not guaranteed at the moment. As illustrated, XML (as an exchange vehicle for the common representation of the data) could close this gap and offers the possibility to exchange data in heterogeneous environments. 1.2. "non-generic" Approach Probably the „worst case scenario“ in terms of reusability is the „non-generic“ approach. This means, that documents are created, but information gets lost: Consider this document. When printed it offers information for human readers, but machines/algorithms will not have a chance to reuse information provided on these pages, because it is not clear, what is the title, the abstract, keywords, and so on. One could guess, that the first sentence written in large bold face could be the title, but is this certain? In another document the author could use another layout and the first „sentence“ is the author of the article. 1.3. The Generic Approach Some Multimedia Authoring tools or dedicated „learning software“ like Lotus Notes/Learning Space offers the possibility to collect data in a structured way. The „generic approach“ means, that information is not only „just entered“ into the system (probably in a layout centric way), but also structured with „meta-information“ information about the information is added. This is an important feature, as reusability of data is rather limited if it is entered in some unstructured way without adding information on content structure, which will be described later on. 2. Persistent Knowledge Acquisition