2012-12-17: Start of this web site: Community Event 2012

History of Med1, Med2, D3, and d3web#

d3web can look back on a long history of development. The first development started about 30 years ago with Med1. Med1 was a diagnostic system for medical purposes. Based on this initial work, several problem solving methods have been developed. A lot of research also focused on how knowledge can be developed easily. All this work has been applied to develop the current version of d3web. Here is a brief overview of the history:

Year of Development Developments...
1983 Development of the first German expert system shell MED1 by Frank Puppe as diploma thesis (MetaEbenenDiagnosesystem 1 (Meta level diagnostic system) with INTERLISP on Siemens BS2000 mainframe computers.
1984 Development of first large expert system based on MED1 by Dr. med. Bernhard Puppe (medical dissertation) about chest pain diagnosis and by Horst Peter Borrmann about car diagnosis (diploma thesis)
1984 Foundation of Inware GmbH by Horst Peter Borrmann and Frank Puppe: Transfer of MED1 to FRANZLIP on UNIX and development of first commercial expert systems: IXMO for Daimler Benz (quality control and production) and SIUX for Siemens (Tuning of data base systems)
1984 - 1987 Development of MED2 as successor of MED1 on FRANZLISP/UNIX (till 1986) and transfer to Common LISP for IBM-AT compatible personal computers (1987)
1987 - 1989 Inware sold 10 commercial licenses of MED2 to various industrial partners. Nevertheless Inware was dissolved in 1989.
1988 - 1989 Development of the graphical knowledge acquisition component CLASSIKA for MED2 by Ute Gappa (diiploma thesis). Therefore a transfer to a LISP dialect was necessary, that supported graphics. Instead of the popular, but expensive LISP-machines at that time, we decided to use the much cheaper Allegro Common LISP which was only available on Apple Macintosh Computers.
1989 - 1994 First funded research project (BMBF) based on MED2/CLASSIKA: „Qualifizierende Arbeitsgestaltung mit tutoriellen Expertensystemen für technische Diagnoseaufgaben“ (Qualifying design of work situation with tutorial expert systems for technical diagnostic tasks). Partners: IBEK business consultancy, Karlsruhe and Voith Sulzer paper machines GmbH.
1991 - 1994 Second funded research project (DFG; project within SFB Artificial Intelligence – Knowledge based systems) „Wissenstransformationen von Diagnosewissen“ (Knowledge transformations of diagnostic knowledge).
1991 Development of a successor of MED2 called D3 (Diagnostic system 3) integrating different knowledge types (heuristic, causal set covering, causal functional, case based knowledge).
1992 - 1994 Transfer of CLASSIKA from Allegro Common LISP to Macintosh Common LISP and transfer and object oriented reimplementation of both D3 and CLASSIKA to CLOS (Common LISP Object System). This implied a new modular architecture as tool box for a configurable diagnostic shell (part of the dissertation of Karsten Poeck). The transfer made it possible afterwards to provide D3/CLASSIKA not only on Macintosh computers, but also on WINDOWS-PCs.
1990 - 1995 Development of a meta-tool for generation of graphical knowledge acquisition component for expert system (dissertation of Ute Gappa). As an application CLASSIKA was reimplemented with the meta-tool. The new graphical knowledge acquisition component allowed a much faster knowledge acquisition mainly by experts themselves. An example was the development of the NEURO-Trainer by the author of the text book Neurologie (neurology), Prof. Poeck, which was distributed as a supplement to the popular text book. The case-based training system was generated from neurological knowledge bases for all diagnoses of the text book. Another example was the development of a large rheumatologica knowledge base by Prof. Schewe, which was later also used for tutorial purposes. This tutorial line of search was continued and resulted finally (2007-2013) in the CaseTrain system (www.casetrain.de), which is widely used now at Würzburg university for providing training cases in nearly all subjects.
1996 Summary of the research and application results with diagnostic expert systems in a text book (Springer): „Wissensbasierte Diagnose- und Informationssysteme – mit Anwendungen des Expertensystem Shell-Baukastens D3 (CD als Beigabe)“ (Knowledge based diagnostic and information systems – with application of the expert system tool box D3 (as additional CD)) by F. Puppe, U. Gappa, K. Poeck and S. Bamberger.
1996 Foundation of a spin off company IISY (intelligente Informationssysteme GmbH; intelligent information systems) of the knowledge bases system group of Würzburg University. One of the founder and still current managing director is Thomas Unglert. IISY was quite successful and first used d3web, then ported it to a new product SOLVATIO and specialized to knowledge based telecommunication service support, where it is currently a market leader.
1995 - 1998 Research project (BMBF): „Kooperierende Diagnostik-Expertensysteme zur Komplexitätsreduktion bei der Entwicklung sehr großer Wissensbasen“ (cooperating diagnostic expert system for complexity reduction of the development of very large knowledge bases) Partners: IuK Dortmund (Insitut für sozialwissenschaftliche Technikforschung; institute for social technologic research, Karlsruhe) and Koenig&Bauer AG, Würzburg. This resulted in the dissertation of Stefan Bamberger „Verteiltes Problemlösen mit wissensbasierten Diagnosesystemen“ (distributed problem solving with knowledge based diagnostic systems). Another result was the definition of diagnostic knowledge acquisition patterns like heuristic decision trees, which resulted from the experiences in various industrial projects with Bosch, Siemens and Koenig&Bauer.
2000 Summary of project results in a text book (Springer): „Wissensbasierte Diagnosesysteme im Service Support – Konzepte und Erfahrungen)“ (Knowledge based diagnostic systems in Service Support – concepts and applications) by F. Puppe, S. Ziegler, U. Martin and J. Hupp.
2001 - 2004 Development of d3web in Java as reimplementation and reorganization of the LISP-based D3 in order to enable use via the web and to provide a modular structure with a kernel engine and plugins. In addition, the graphical knowledge acquisition component CLASSIKA was reimplemented in Java too, named KnowME (Knowledge Modelling Environment). In his dissertation Joachim Baumeister developed concepts for anomaly detection, refactoring and automatic testing thus supporting an agile development style for knowledge acquisition.
2005 - 2008 Starting with several student projects for knowledge acquisition with standard office tools (text and calculating systems) and the diploma thesis of Marcus Friedrich, this approach was taken further to a wiki-based Knowledge acquisition tool called KnowWE (Knowledge Wiki Environment). It was first implemented with the wiki system TWIKI (Pearl based wiki), but later on ported to use the Java based JSPWiki system. Besides enabling distributed knowledge acquisition, a major feature of KnowWE was the smooth integration of formal and informal knowledge.
2009 Foundation of denkbares GmbH by Volker Belli, which uses and develops d3web/KnowWE as an open source tool offering management and knowledge engineering support in large projects.
2009 - 2011 denkbares stabilizes the code of d3web/KnowWE and presents regular releases beginning in 2009. Development of several large commercial applications with Dräger, Umweltbundesamt, Krauss-Maffei Wegmann, and Digitalis of d3web/KnowWE.
2011 - 2014 Joint project WimVent: Wissens- und modellbasierte Therapiestuerung in der Medizin, von der komplizierten Beatmungseinstellung zur individuellen strategischen Therapieführung (by BMBF) Partners of Würzburg University: Dräger AG, Univ. Kiel, Univ. Freiburg, FH Furtwangen. Part of the project is a graphical guideline language and tool Diaflux based on KnowWE.
2012 - 2014 Joint project JuriSearch: Entwicklung eines wissensbasierten Systems mit interaktiver Benutzeroberfläche zur Strukturierung und Transparenzsteigerung für juristische Problemstellungen (by FuE Programm „Informations- und Kommunikationstechnik“ des Freistaates Bayern), Partners of Würzburg University: RenoStar (Großwallstadt). Part of the project is a WUMPS, a web upload and parser for tagged office documents for acquistion of large d3web knowedge bases by experts and ProKEt, a consultation dialogue prototyping tool.
2013 - 2014 Autonomes Diagnosesystem für Satelliten (by DLR – Deutsches Zentrum für Luft- und Raumfahrt), Würzburg University.