cassel@gaia.sce.carleton.ca (Ron Casselman) (06/27/91)
Has anyone had any experiences with the GRID specification for large
systems?
Harold Ossher the inventor of GRID publicized this technique for
specifying the structure of systems in two articles.
@article{ossher-89,
author = {H. L. Ossher},
journal = {IEEE Transaction on Software Engineering},
number = {11},
pages = {1397-1416},
title = {A Case Study in Structure Specification: A Grid Description of Scribe},
volume = {15},
year = {November 1989},
annotate = "The grid mechanism is a graphical notation
for describing the structure of large software systems and for
specifying and enforcing structuring the structuring disciplines. It
is intended to present complex structures in a clear and intuitive
manner, yet it is formal: consistency between a system and a grid
specification can be checked automatically."
}
@inbook{ossher-87,
author = {H. L. Ossher},
address = {Cambridge, MA},
booktitle = {Research Directions in Object-Oriented Programming},
editor = {B. Shriver and P. Wegner},
pages = {219-252},
publisher = {MIT Press},
title = {A Mechanism for Specifying the Structure of Large Layered Systems},
year = {1987},
annotate = "The grid specification applied to object-oriented systems."
}
It seems interesting because it allows one to specify formally
relationships between components that are usually managed at the
implementation level using programming conventions and engineering
discipline.
Is anyone aware of similar techniques?
Do these techniques scale to really large systems--say 5 million
lines of code and up?
Does anyone know of tool to support GRID. I understand that Dr. Ossher
was developing such a tool at the IBM Thomas J. Watson Research Center?
DOes anyone know where I can connect Dr. Ossher - IBM says he is no
longer an employee.
--
Ron Casselman (613) 788-5726 Systems and Computer Engineering,
uunet!mitel!sce!cassel (uucp) Carleton University,
cassel@sce.carleton.ca (internet) Ottawa, Ontario, Canada K1S 5B6.