Infrastructure for the representation and electronic exchange of design knowledge

This paper develops the concept of knowledge and its exchange using Semantic Web technologies. It points out that knowledge is more than information because it embodies the meaning, that is to say semantic and context. These characteristics will influence our approach to represent and to treat the knowledge. In order to be adopted, the developed system needs to be simple and to use standards. The goal of the paper is to find standards to model knowledge and exchange it with an other person. Therefore, we propose to model knowledge using UML models to show a graphical representation and to exchange it with XML to ensure the portability at low cost. We introduce the concept of ontology for organizing knowledge and for facilitating the knowledge exchange. Proposals have been tested by implementing an application on the design knowledge of a pen.

💡 Research Summary

The paper presents an infrastructure for representing and electronically exchanging design knowledge, emphasizing that knowledge is more than raw information because it carries meaning and context. To make knowledge both human‑readable and machine‑processable, the authors combine three standard technologies: UML for graphical modeling, XMI (the UML serialization format) for portable exchange, and RDF/RDFS for semantic description.

First, design knowledge is modeled with UML. Static aspects such as component hierarchies and ontological concepts are captured in class diagrams, while dynamic behavior (processes, interactions, state changes) is expressed with sequence, activity, collaboration, and state‑chart diagrams. The use of OCL (Object Constraint Language) allows precise definition of constraints and invariants, reducing ambiguity and enabling automated validation.

Second, the UML models are serialized to XMI, an XML‑based interchange format defined by the OMG. XMI guarantees that the same model can be imported into different CAD, PDM, or knowledge‑management tools without loss of fidelity, thereby satisfying the portability requirement.

Third, a lightweight ontology is built using RDF and its schema language RDFS. The ontology provides a controlled vocabulary and explicit relationships (e.g., a “composition” property that denotes strong aggregation between resources). By linking each knowledge card to RDF descriptors, the system enables semantic queries via SPARQL, facilitates automatic reasoning, and ensures that all participants share a common understanding of terminology. The authors also incorporate Dublin Core metadata to describe documents and provenance.

The methodology is illustrated with a case study on the design of an “advertising pen.” Knowledge extracted from the CYGMA knowledge book (components, parameters, constraints, etc.) is first modeled in UML, then exported as XMI, and finally annotated with RDF/RDFS. A three‑tier web architecture (client, application server, storage) supports two user modes: knowledge viewing and knowledge capture. In capture mode, a remote designer creates a structured knowledge card in XML; the server stores the card simultaneously in a relational database (for quick retrieval) and in the RDF store (for semantic access). In viewing mode, users can retrieve specific parts of the pen design by issuing SPARQL queries, obtaining both the graphical model and the associated semantic metadata.

The paper compares the proposed approach with ISO STEP, noting that STEP focuses on product data from a manufacturing perspective and lacks support for organizational or abstract design concepts. UML, by contrast, can represent multiple viewpoints through its diverse diagram types, making it more suitable for collaborative design environments.

Key advantages of the proposed infrastructure are: (1) reliance on widely accepted standards (UML, XMI, RDF/RDFS) which lowers adoption cost and eases integration with existing tools; (2) a hybrid representation that combines visual diagrams, textual descriptions, and formal schemas, improving user comprehension and reducing ambiguity; (3) a web‑based service layer that enables distributed teams to share and synchronize knowledge while preserving semantic consistency.

Limitations acknowledged include UML’s inability to capture every nuance of design knowledge without supplemental textual documentation, and RDF/RDFS’s limited expressive power for complex logical inference compared to more heavyweight ontology languages. The authors suggest future work involving OWL ontologies and advanced reasoning engines to enhance automated knowledge generation and validation, performance testing on large‑scale design projects, richer user interfaces for non‑expert contributors, and the addition of security and version‑control mechanisms to meet enterprise requirements.