Greg Nordstrom

Smart Dust: communicating with a cubic-millimeter computer

W hat do Rational Rose, Simulink, and LabVIEW have in common? At first, these tools seem very different. Rational Rose (http://www.rational.com) is a visual modeling tool, Simulink (http:// www.mathworks.com) is a hierarchical block-diagram design and simulation tool, and LabVIEW (http:// www.ni.com) is a graphical programming development environment. Despite the different terminology, these three tools share a common underlying theme: Each is an integrated set of modeling, model analysis, simulation, and code-generation tools that help design and implement computer-based systems (CBSs) in a specific, well-defined engineering field. These tools and other popular domain-specific integrated development environments can help capture specifications in the form of domain models. They also support the design process by automating analysis and simulating essential system behavior. In addition, they can automatically generate, configure, and integrate target application components. These environments translate the verified design—expressed in a domainspecific, primarily graphical modeling formalism—into a variety of artifacts that constitute a CBS implementation. These artifacts can include glue code, database schema, and configuration tables. These tools use domain-specific modeling languages that allow developers to represent essential design views and to both formally express and automatically enforce integrity constraints. These tools also support model composition that is synergistic with the design process in the particular engineering domain. Other benefits include having integrated models as opposed to relying merely on source code. In addition, the common input—that is, the shared design model—guarantees the consistency of different analysis results as long as all of the applied generators are correct. While the industry understands the welldocumented benefits of domain-specific, integrated modeling, analysis, and application-generation environments, their high cost represents a significant block to wide acceptance and application. Consequently, these tools are available only for domains with large markets in which high volume offsets the substantial initial investment cost. For CBSs in smaller, specialized domains, or even for single projects, the industry needs technology that can help rapidly and efficiently compose these environments from reusable components.Domain-specific integrated development environments can help capture specifications in the form of domain models. These tools support the design process by automating analysis and simulating essential system behavior. In addition, they can automatically generate, configure, and integrate target application components. The high cost of developing domain-specific, integrated modeling, analysis, and application-generation environments prevents their penetration into narrower engineering fields that have limited user bases. Model-integrated computing (MIC), an approach to model-based engineering that helps compose domain-specific design environments rapidly and cost effectively, is particularly relevant for specialized computer-based systems domains-perhaps even single projects. The authors describe how MIC provides a way to compose such environments cost effectively and rapidly by using a metalevel architecture to specify the domain-specific modeling language and integrity constraints. They also discuss the toolset that implements MIC and describe a practical application in which using the technology in a tool environment for the process industry led to significant reductions in development and maintenance costs.

Metaprogrammable toolkit for model-integrated computing

Model-integrated computing, specifically model-integrated program synthesis (MIPS) environments that include visual model building, constraint management, and automatic program synthesis components, are well suited for the design and implementation of complex computer based systems. However, building such an environment from scratch for each new domain can be cost-prohibitive. This paper presents a toolkit that makes the rapid creation of MIPS environments possible through metaprogramming.

The new metamodeling generation

Model integrated computing (MIC) is an effective and efficient method for developing, maintaining, and evolving large-scale, domain-specific software applications for computer-based systems (CBSs). On a higher level, it is possible to use MIC to develop, maintain, and evolve the meta-level tools (metamodeling environments) themselves, by modeling the metamodeling environment (meta-metamodeling). This paper documents the evolution of one metamodeling environment into another: specifically the design choices of the newer metamodeling environment with regard to the old one, and the solutions to problems that were introduced with the change.

Modeling of supply chain: a multi-agent approach

Describes the development of an agent-based software system for assisting in decision-making regarding supply chain management and the efficient and effective use of electronic data interchange (EDI) in the automobile industry. Such a system can be applied to different types of industries with some domain-specific modifications. The core architecture is built around the concept of a software-based agent that is programmed internally to interact with other external agents in a pre-defined manner. We are developing a MIC (model-integrated computing) based supply chain management modeling environment. This environment allows domain experts to create models of the software agents to simulate and control the actual online negotiation processes. The modeling environment allows modeling of agent behavior, as well as defining agent-to-agent interaction scenarios.

A graduate-level course on CBS design tool development

A graduate-level course addressing the design and development of modeling tools for computer-based systems (CBS) is described. A model integrated computing (MIC) approach to the design of CBS is taught wherein domain applications are automatically generated and/or configured from domain-specific models. Students are taught to recognize and comprehend the tenets of CBS design and implementation in various engineering environments. Both instructor- and student-led lectures are used to develop an understanding of systems engineering processes, modeling tools, techniques and methodologies and their place in the development of CBS. Homework assignments are used to enhance the students ability to apply MIC to CBS development and design. To complete the course, students are required to design and develop a CBS modeling tool for use in a real-world engineering domain. Students demonstrate their comprehension of MIC and CBS concepts through homework, quizzes, and in-class presentation of their proposed project and a final project presentation including a demonstration. Students leave the course with a strong foundation in both CBS tool design and MIC.

Towards a standard for model specification and storage

Software production has become an industrial task usually involving teams of programmers working on complex problems to produce large, even huge software systems. Globally distributed teams are doing a growing share of all software development work. The management of software engineering teamwork, especially of a temporally and/or spatially distributed team, presents an enormous organizational challenge as well as an intricate technical problem, as such distributed teamwork requires tool support for coordination of cooperative activities, maintenance of project control, and sharing of information. Domain-specific Model Integrated Program Synthesis environments are created according to a modeling paradigm: a description of the class of models that can be created using the system. Just as model integrated computing applications are executable instances of domain models, domain models can be viewed as instances of metamodels. The representation of these models and the modeling paradigm is unique to the specific modeling environment. This poses a major problem for portability of models from one modeling environment to another. The purpose of the paper is to explore the possibility of a common standard for the storage of models, in what framework the standard should exist, and who should define the standard.