Nine month progress report submitted for continuation towards a PhD
Toward a Canonical Method to Solve Patterns of Ontology Modelling Issues 8 ____________________________________________________________________________________________________________The second ontology is the ReSIST Ontology which is still being developed at present. This ontology models the domain of resilient computing. The third ontology is the ReSIST University Curricula Ontology. It models the concept of a university course related to the topic of resilient computing. The following three sections describe the last two ontologies mentioned and a series of problems encountered during the modelling phase that contributed in situating ontology modelling as the main research focus of this report. 3.1.2. An Ontology for Resilience and Survivability in IST An exploration of existing ontologies was conducted with the goal of identifying a candidate that could be re-used for our target domain. This task was performed by searching in Swoogle1 (an indexing and retrieval system of ontologies for the Semantic Web) for key concepts from our target domain such as: resilience, dependable, dependability, fault, fault-tolerant, fault-tolerance, etc. The same searches were issued in Google instructing to check only for ontology files, (by using the engine’s filetype:owl web search feature). Both tools retrieved several OWL files, (although in the case of Swoogle the number of results was larger and more relevant), but none of them was a suitable option mainly because they applied to different domains. The closest candidate was an ontology for the ACM Computing Classification System (ACM 2002). However, this classification scheme is intended for the whole field of Computing, and does not provide the sufficient level of detail for our needs. An observation drawn after going through the experience of searching existing ontologies as described is that it is not very intuitive to understand from the OWL files found, the purpose or domain that they belong to. It’s not clear if this is due to the lack of relevant comments in the files themselves or if something could be done on how these search engines present their results to the user. At the view of the results, it was concluded that the ontology for ReSIST will have to be built from scratch. The first step when building an ontology from scratch is to acquire the knowledge of the domain to model (Noy and McGuinness, 2001)(Fernandez-Lopez et al., 1997)(Uschold and King, 1995)(Gruninger and Fox, 1995). Usually this process requires contacting experts in the domain that can facilitate this information and can be very time-consuming. For our purpose it was agreed to use the terms defined in a relevant key publication in the field as the initial glossary of terms baseline to model in the ReSIST ontology (Avizienis et al., 2005). This would alleviate dramatically the knowledge acquisition task.1http://swoogle.umbc.edu/