Abstract
Today’s organizations have well developed systems for capturing financial data and producing reliable and accurate information. An area that is often overlooked is the importance of knowledge held by individuals associated with the organization. This is also true in an educational setting, where learners can often contribute real-world and personal experiences.
The difficulty of capturing this knowledge provides many challenges, in both business and educational settings. In order to be of any use, this knowledge has to be captured at its source and easily disseminated among those who will be interested or affected by this knowledge. The World Wide Web may be a potential technology platform.
The research in progress looks at whether this data/information can be effectively and efficiently captured, managed and retrieved, in an educational context.
Over the past 10 years my teaching resources have evolved from electronic file sharing to an on-line environment. HTML linked documents in a networked file system allowed for easy distribution of teaching materials and has been the method of choice up to now. The availability of an interactive web server with database support meant that the content could be expanded to use this technology. Using a database approach allows issues that were not well addressed in the file-sharing structure to be investigated. These include: problems of keeping the material current; search facilities and the capturing of student comments and knowledge.
A prototype system has been developed and selected existing content was re-engineered into this format. I have used the system to deliver teaching content to students studying database, multimedia and web design.
From my observations in an educational setting, I have found that the students can provide valuable resources in one of the two ways. Firstly, students can contribute from their own "real-world" experiences. For example, while teaching coding systems a student who was also a bank employee discussed how account numbers were codified. Secondly, students are often researching the topic whether in a learner directed mode or discovery learning mode. This contribution to the topic knowledge base is often held by the individual and most often lost to their peers.
A second and equally important part the students can play is that of moderation of resources. In one instance recently, a resource indicated that a "modern computer that would support multimedia was a 486 with a large 20MB Hard Drive"! It turned out that the resources became outdated without anyone noticing it in the vast amount of material.
The issue here is how to capture the knowledge and experience from the students, and from others who happen to be there, and place it into the current context. Typically in a web framework this is done via e-mail or through bulletin-boards. In both of these cases the knowledge is separated from the content and time and effort is required to join the content to the knowledge, and, unfortunately often this does not happen.
A content management system is being developed to resolve this problem. A database approach was used and content was normalized into a subject domain. In order to maintain some meaning and structure a backbone taxonomy was created, based on the work of Guarino & Welty (2002).
Since the original approach had all the information/knowledge in HTML linked documents, the first task was to develop an entity that would reproduce the static pages. The concept of a small logical piece of knowledge evolved. In some systems (for example, Hyperskript) this is known as a fragment but to enable a progressive definition, the term sniplet has been coined. Since ‘Learning Object’ is a pre-defined term, which has a rather broader meaning, it was not seen as replacing the term “Sniplet”. For example, the IEEE (1999) definition of Learning Objects includes learning objectives, persons, organizations, or events.
From the original file sharing system two major requirements emerged to enable content delivery. Firstly, the ability to provide detailed content in the form of handouts, suitable for a printed format, and second to provide a summary suitable for an overhead projector/data-show. After several prototypes, the content fragment that proved to be most workable was a piece of knowledge or information that could be represented by one overhead transparency, and in order to provide a way to refer to this, the term “Sniplet” was coined.
Core attributes of the sniplet are:
Sniplet (Creator, id, Backbone taxonomy id, title, description, summary, multimedia id and bibliographic id)
Based on this entity, other related entities were created such as backbone taxonomy, bibliography, and multimedia elements.
Figure 1: Sample Sniplet
Figure 2: Sniplet Architecture
Once a workable schema was in place, the next task was to enable students to provide input in the system during the learning process. It was seen vital that the capture of knowledge had to happen within the domain that the students were focusing on (highly contextual), and should represent that context correctly to others.
Principally, two types of knowledge need to be captured. The first is as an annotation to existing content. The techniques of annotation have been found useful to retain context while avoiding unnecessary changes in the original knowledge object (in our case, a Sniplet). On-line news bulletins, such as www.techcentralstation.com, allow discussion treads, user feedback and user ratings directly attached to the article. This technique has been adopted in the sniplet database prototype. Secondly, new content needs to be added to the knowledge base. To maintain the integrity of the knowledge base, only users with suitable access rights should be able to add this level of content, or additions need to be moderated and accepted prior to addition in the knowledge base.
The final part of the content management system is the ability to reorganize the domain content for use in different teaching and learning situations. At this level it will be important that the annotations and any new or additional content are flagged so that any anomalies or updated content can be added to this reorganized taxonomy.
The system is work in progress and can be viewed at http://www25.brinkster.com/verhaart/.
Here are some issues to consider:
Guarino,N & Welty, C. (2002, Feb) Evaluating Ontological Decisions with OntoClean. Communications of the ACM, 45 (2), 61-65.
IEEE Learning Technology Standards Committee (LTSC) (1999, Sep 5) Learning Object Metadata, http://ltsc.ieee.org/doc/wg12/LOM3.6.html.
