D3web
| Company type | Open-source software |
|---|---|
| Industry | Software development, Knowledge engineering |
| Founded | 1980s |
| Founder | Prof. Dr. Frank Puppe |
| Headquarters | Germany |
| Products | Knowledge-based systems framework, diagnostic engine, KnowWE, KnowME, Dialog2 |
| Website | sourceforge |
d3web is an open-source and cost-free framework designed for knowledge-based systems (expert systems), developed using Java, with knowledge representation handled through XML and Office-compatible formats. The entire suite of components is distributed under the Lesser General Public Licence (LGPL).
The diagnostic engine of d3web facilitates reasoning and data persistence for problem-solving knowledge, encompassing methods such as decision trees, heuristic rule sets, set-covering approaches, and diagnostic flowcharts. It supports integration into external software solutions either as embedded modules or OEM products, with multiple ready-to-use components available.
Components
d3web provides a modular platform tailored for both the creation and execution of problem-solving knowledge:
- KnowWE (Knowledge Wiki Environment): This semantic wiki, built atop JSPWiki, allows users to author and run problem-solving knowledge through a wiki interface. Knowledge bases created here can be exported for embedded or OEM applications, featuring compatibility with knowledge exchange via OWL ontologies.
- KnowME (Knowledge Modelling Environment): A desktop rich-client tool previously used for developing d3web knowledge bases. Although it has been superseded by KnowWE and is no longer under active development, it enabled both creation and execution of problem-solving knowledge for embedded or OEM deployment.
- Dialog2: A web-based demonstration application illustrating the core reasoning capabilities of d3web. Developed using Java Server Faces, it functions as a ready-to-use solution and as a base for customized knowledge-based interview systems.
Application Domains
d3web has been utilized in numerous academic and industrial projects, especially in the following fields:
- Medical diagnostics, documentation, and treatment planning
- Technical fault diagnosis
- Surveillance of technical equipment
Prominent applications built upon the d3web diagnostic core include:
- SmartCare(c) by Dräger: a medical closed-loop system designed to assist in weaning patients from mechanical ventilation
- SonoConsult: a medical assistance tool for evaluating sonographic examinations (available only in German)
- eDOC: an online self-diagnosis platform addressing various medical conditions (German language only)
History
The origins of d3web trace back to research led by Prof. Dr. Frank Puppe at the University of Würzburg, Germany, during the 1980s, starting with the development of medical expert systems MED1 and MED2. While initially concentrated on medical diagnosis, this methodology was expanded by its successor D3 to encompass a wider range of diagnostic tasks. In contrast to its LISP-based predecessors, d3web represents a full Java reimplementation, improving integration capabilities and scalability suited to contemporary business settings.
See also
Related systems
- CLIPS: a public domain expert system development environment.
- ILOG rules: a system managing business rules.
- JBoss Drools: a business rule management system (BRMS).
- JESS: a Java rule engine that extends CLIPS functionality.
- Prolog: a versatile logic programming language.
- DTRules: an open-source Java rule engine based on decision tables.