A meta-language for system architecting: A case study based on NASA's Apollo Program

Thursday, March 17, 2005 - 7:00pm
Benjamin Koo and Jay Conne

Language is our basic tool of thinking and communicating. A meta-language supports the consistent creation of domain-specific languages, such as the languages of chemistry, electrical engineering or rocket science. And having an executable meta-language gives you an automated language generation and computation capability.

Ben Koo's recent research developed an executable meta-language that supports system architects' modeling process by automating certain model construction, manipulation and simulation tasks. This language facilitates systematically communicating architects' intent with a wide range of stakeholders and to organize knowledge from various domains. The investigation into existing architecting approaches and technologies has pointed out the need to develop a simple, yet formal language that expresses multiple layers of abstractions, provides reflexive knowledge about the models, mechanizes data exchange and manipulation, while allow integration with legacy infrastructures. A small set of linguistic primitives, stateful objects and processes that transform them, were identified as both required and sufficient building blocks of the meta-language, specified as an Object-Process Network (OPN).

To demonstrate OPN's practical value in large-scale engineering systems, the research applied OPN to two space exploration programs and one aircraft design problem. In these experiments, OPN was able to significantly simplify the modeling and architectural reasoning process by automating a number of manual model construction, manipulation, and simulation tasks.

Lecturer Biography: 

Dr. Benjamin Koo served as a software architect in the Information Technology industry for many years. He provided enterprise application integration services to various Fortune 500 corporations, mostly in the Financial and Telecommunication industries. He received his doctoral degree from MIT, where his primary research interest focused on the theory and instrumentation for system architecting. He is currently a Post-Doctoral Associate at MIT, working on the Concept Evaluation and Refinement program sponsored by NASA's Space Exploration Program.

Jay Conne has over four decades experience in communications, systems and data engineering, as well as training, and business development. He has served in these roles in the Bell System, General Electric, Honeywell, Burroughs Corp, Digital Equipment Corp. and for financial services, transportation and consulting firms. Jay is currently working with Ben Koo on productizing OPN. Jay is a former GBC/ACM President and the current Chapter VP.