The basic ideas for Tom Runge's dissertation were taken, in his own words, from a visionary paper by Bill Gear, his dissertation advisor. A special class of object-oriented modeling methodology, the bond graph, had been around even before then. The bond graph approach to physical system modeling was conceptualized by Hank Paynter on April 24, 1959 who, in turn, claims that his work had been influenced in major ways by the even earlier work of Gabriel Kron.
Yet, the two dissertations by Runge and Elmqvist were far ahead of their times. As a consequence, neither technology received a high degree of attention for close to a decade. The computer hardware of the late seventies was far from ready for the proposed software technologies (Hilding once built his own mouse interface, because he couldn't find one for the VAX), and the market demand for such sophisticated modeling software was simply not there yet.
Only the early nineties, with their growing demands for simulation models of ever increasing complexity describing models of advanced engineering systems with component models stemming from different application fields, as they are used in e.g. mechatronics, electronic packaging (with electronic, mechanical, and thermal submodels), aircraft electronics (subjected to high temperature, strong magnetic fields, and heavy mechanical vibrations), the electronic fuel injection control system of a car, to mention just a few examples, called for the sophistication that object-oriented modeling tools have to offer. This growing demand went hand-in-hand with the deployment of graphical workstations that finally provided a platform powerful enough to host a sophisticated object-oriented modeling environment.
The object-oriented physical systems modeling tools available today all offer object-diagram editors as well as textual model representations. They all support a mixture of symbolic model preprocessing with advanced DAE solvers. Had the theses of Runge and Elmqvist been viewed by many as somewhat esoteric at the time they were written, no-one questions any longer the relevance and significance of this research branch. In fact, every one of the above software tools is dramatically better than anything that had been on the market earlier in terms of general-purpose physical system modeling support.
This panel discussion is a first attempt at bringing many the major players together and have them discuss, with inclusion of the attendees, what their views are w.r.t. future developments in this exciting and dynamic research area.
Interested in reading the full paper? (2 pages, 73,840 bytes, pdf)