Review of Ontologies for sharing, Ontologies for use
This article is well written and structured, stressing the point that in order to achieve the maximum usability and re-purposing of knowledge contained in an ontology, the ontology should be constructed in small manageable modules with a clearly delinated scope and follow normalisation principles. This reviewer agrees with this principle of ontology design however, has the following comments:
One of the main advantages of producing modular ontologies is that not only can they be extended and constrained for a new purpose, they can also be easily integrated with other modules to produce a new, or more comprehensive ontology. Therefore more space could be given towards the benefits of modular ontology design to ontology integration.
In order to re-use or integrate ontology modules, the process will be considerably easier if the modules are built with the same design principles, such as a common set of relations of the use of an upper ontology. The paper should describe the relevance of upper ontologies and common relations, in respect to their design principles.
It is not clear that these principles are applicable to all modular ontology development (i.e Reference and application) should it be a restriction on application ontologies that they should be shared?
The article presents principles that could be applied from day one of development of a modular ontology, whereas this is the ideal situation if we were starting ontology development again, some consideration should be given to how we make use of the ontologies that exist to be more like modular ontologies. For example, identifying overlap between these ontologies, and creating modular ontologies, applying the stated principles from these overlapping terms, which could then be re-imported back into the original ontologies.
I dont believe the “world” or the “Semantic Web” as a whole can agree on a singular view of the world in one go. The development of modular ontologies, motivated by identifying overlap between existing ontologies, seems more appropriate on a domain by domain basis – and then on a domain domain overlap. The example of dc:title is given, However, a title is a prefix or suffix added to a person’s name to signify either veneration, an official position or a professional or academic qualification. Already we have a conflict. In the domain of publishing it is clear what a title refers to, in the wider world or semantic web it is not.