Abstract
We are reporting on our experience in building and deploying OWL ontologies from within a newly introduced PSU ontology repository (POR) into a number of application software. These ontologies are also exposed to public outsiders to edit and extend (Open) and allow access to similar Internet ontologies using public ontology search engines such as SWOOGLE and IBM WATSON. This environment is mainly used to empower software application developers with semantics (analogues to Semantic Web) to limit development complexity, provide knowledge management, and organize software project information. OWL ontologies are machine readable; they can be incorporated into the software application development as accessible knowledge base shared resources which parts of the software can access as needed knowledge units via regular protege programming interfaces: Script Tab, and Java API. These shared resources reduce development complexity, and promote re-use through linking similar and cross domains ontologies. We have been using and testing them within software applications to model certain artifacts (such as context and profile management, Internet of Things, sensor networks, cinema production, and risk and security management components) during the application software development life cycle. In this paper we report on our experience with each of these software applications: their artifacts ontology structure, representation and engineering, validation and verification, development and testing, and ontology querying and usage. We introduce a new framework of mixing "ontology engineering and access" with "application software development environment". In order to measure the gains achieved through this proposed framework, we ran a "Questionnaire" among developers to evaluate - What are the benefits gained of mixing "Ontology Engineering and Access" with "Application Software Development Environments"? - Is the "Framework" presented perceived useful by developers? - Are the "Consolidated Customized" ontologies constructed using the framework better, in some modeling quality sense, than the ontologies constructed without the framework? - Are the tasks given to developers done faster when using the framework? - How do developers use the framework provided, and what support would be beneficial? Subjective opinions show that 2/3 of developers perceive the framework as useful or very useful. We found out that developers feel that they have constructed better quality ontologies and that they are "guided" by the framework. As such the main purpose of the open POR to facilitate inhouse and out-of-house resources for building consolidated customized ontologies has been fulfilled.