This paper by Thomas R. Gruber introduces a translation approach to making ontologies portable across different representation systems. An ontology is a specification of a conceptualization, defining classes, relations, functions, and other objects in a shared domain of discourse. The paper discusses the importance of portability in supporting the sharing and reuse of formally represented knowledge among AI systems. It addresses technical challenges such as accommodating stylistic and organizational differences among representations while preserving declarative content, and translating from expressive languages into restricted languages while maintaining computational efficiency. The paper describes Ontolingua, a system that translates ontologies written in a standard, system-independent form (Knowledge Interchange Format, KIF) into specialized representations, including frame-based systems and relational languages. Ontolingua translates definitions into forms suitable for implemented representation systems, allowing researchers to share and reuse ontologies while retaining the computational benefits of specialized implementations. The paper also discusses the role of ontologies in knowledge sharing, emphasizing the need for conventions at three levels: representation language format, agent communication protocol, and specification of content-specific agreements. It provides an example of an ontology for bibliographic information and explains how Ontolingua translates definitions into different representation systems, such as Epikit, Loom, and KEE. The design of Ontolingua balances expressiveness, system independence, and ease of extension. It recognizes common representation idioms and transforms them into a canonical form, which is then translated into target systems. The Frame Ontology, an ontology of domain-independent representational idioms, is used to define the vocabulary and constraints for ontologies written in Ontolingua. The paper concludes by discussing the limitations of Ontolingua, noting that it does not support query processing and is inherently incomplete with respect to KIF. However, it provides a flexible and adaptable tool for making ontologies portable across different representation systems.This paper by Thomas R. Gruber introduces a translation approach to making ontologies portable across different representation systems. An ontology is a specification of a conceptualization, defining classes, relations, functions, and other objects in a shared domain of discourse. The paper discusses the importance of portability in supporting the sharing and reuse of formally represented knowledge among AI systems. It addresses technical challenges such as accommodating stylistic and organizational differences among representations while preserving declarative content, and translating from expressive languages into restricted languages while maintaining computational efficiency. The paper describes Ontolingua, a system that translates ontologies written in a standard, system-independent form (Knowledge Interchange Format, KIF) into specialized representations, including frame-based systems and relational languages. Ontolingua translates definitions into forms suitable for implemented representation systems, allowing researchers to share and reuse ontologies while retaining the computational benefits of specialized implementations. The paper also discusses the role of ontologies in knowledge sharing, emphasizing the need for conventions at three levels: representation language format, agent communication protocol, and specification of content-specific agreements. It provides an example of an ontology for bibliographic information and explains how Ontolingua translates definitions into different representation systems, such as Epikit, Loom, and KEE. The design of Ontolingua balances expressiveness, system independence, and ease of extension. It recognizes common representation idioms and transforms them into a canonical form, which is then translated into target systems. The Frame Ontology, an ontology of domain-independent representational idioms, is used to define the vocabulary and constraints for ontologies written in Ontolingua. The paper concludes by discussing the limitations of Ontolingua, noting that it does not support query processing and is inherently incomplete with respect to KIF. However, it provides a flexible and adaptable tool for making ontologies portable across different representation systems.