In this study we present ontologies as a potential “silver bullet”, enabling automated knowledge sharing and reuse among both human and computer agents, because of their ability to interweave human and machine understanding through formal and real-world semantics. Because there are so many different types of ontologies, ranging from simple word lists to comprehensive ontologies with the expressive power of full first-order logic, we identify two dimensions used for the classification of ontologies. Ontologies can be classified according to their generality and according to their expressiveness. With generality we mean the breadth of the ontology. Some ontologies try to capture all terms in natural language, while other ontologies are very specific to certain domains or certain applications. The expressiveness of an ontology relates to the degree of explication of the (meta-)knowledge, which is captured in the ontology. When more relations and more constraints are captured in the ontology, the ontology becomes more expressive, since it captures the knowledge of the domain on a more detailed level. The languages used for specifying an ontology restricts the expressiveness of the ontology as well as its usability and reusability across applications. We describe several initial languages, which have been developed originally in the context of Knowledge Representation systems, for describing ontologies, as well as the development of current ontology languages for the Semantic Web. Because of the complexity of the task and the many demands on ontologies in terms of usability and reusability, many ontology engineering methodologies have been developed. Many of these engineering methodologies use the design principles we identify as being important for the design of good ontologies. We evaluate some of these methodologies in order to identify how the problems facing the development of ontologies are being addressed. Ontologies have the potential of enabling true knowledge sharing and reuse among heterogeneous agents, both human and computer. There are, however, some obstacles that still need to be taken, in the form of ontology language and ontology engineering issues, which are described in this report. Other prerequisites for the usage of ontologies on the Semantic Web include the development of intelligent agents and the actual annotation of data sources on the current World Wide Web using ontologies. One major open challenge is still the alignment of different ontologies in order to allow inter-operation between heterogeneous agents. These prerequisites are not addressed in this study, but there are many research efforts under way, which address these issues.