In its introduction, this article gives a short state of the art about ontologies of knowledge representation languages (KRLs) and the problems caused by i) the lack of relations between these ontologies, and ii) the lack of ontologies about notations (concrete syntaxes). For programmers, these are the difficulties of importing, exporting or translating between KRLs; for end-users, the difficulties of adapting, extending or mixing notations. To show how these problems can be solved, this article first shows how concepts of the main KRL standards can be aligned and organized. Then, it shows how this KRL model ontology can be reused and completed by a notation ontology. Based on these two ontologies, KRLs models and notationsand thereby parsing and generation-can be specified in a concise way that even KRL end-users can adapt. The article gives representative examples. For these ontologies or specifications, a concise KRL notation is introduced and used. However, the presented approach is independent of any notation and model that has at least OWL-2 expressiveness. Thus, the results can easily be replicated. A Web address for the full specification of the two ontologies, and for a knowledge server to test or use them, is also given.

No abstract

This article introduces KRLO, an ontology of knowledge representation languages (KRLs), the first to represent KRL abstract models in a uniform way and the first to represent KRL notations, aka concrete models. More precisely, this article illustrates the content, principles and kinds of use cases for such an ontology. One kind is to help design tools handling many KRLs, and hence parsing, semantically analyzing and exporting knowledge expressed in these KRLs. Another kind is to let the end-users of these tools design or adapt KRLs. This ontology also supports translations based on equivalence or implication relations between types as well as some structural translations. They can be exploited by inference engines handling the expressiveness of RIF-BLD, i.e., of Datalog like rules.

Abstract-This article proposes an ontology design pattern leading knowledge providers to represent knowledge in more normalized, precise and inter-related ways, hence in ways that help automatic matching and exploitation of knowledge from different sources. This pattern is also a knowledge sharing best practice that is domain and language independent. It can be used as a criteria for measuring the quality of an ontology. This pattern is: "using binary relation types directly derived from concept types, especially role types or process types". The article explains this pattern and relates it to other ones, thereby illustrating ways to organize such patterns. It also provides a top-level ontology for generating relation types from concept types, e.g., those from lexical ontologies such as those derived from the WordNet lexical database. This generation and categorization helps normalizing knowledge, reduces having to introduce new relation types and helps keeping all the types organized.

Via a comparison of the currently used language-based components for knowledge sharing, this article first highlights the difficulties caused by the inexistence-and hence absence of exploitation-of a shared core ontology of knowledge representation languages (KRLs), i.e., i) an ontology of KRL abstract models which represents, aligns and extends standards, and ii) an ontology of KRL notations. For programmers, these are the difficulties of importing, exporting or translating between KRLs; for end-users, the difficulties of adapting, extending or mixing notations. This article then shows how we have built this shared core ontology plus a tool for exploiting it. We use them for specifying, parsing and translating KRLs, thus allowing their use without additional programming. This ontology can be represented in any KRL that has at least OWL-2 expressiveness. Thus, the results can easily be replicated. A Web address for the tool and the full specifications is given.