Geoffrey Darnton

Bournemouth

Senior Academic

This sounds like the re-birth of a project started in the 1960s - the ISDOS project (Information System Design and Optimization System) at the University of Michigan under Prof. Daniel Teichroew. The goal was to represent the overall system requirements with sufficient information about requirements and details, so that those parts that can be automated can be reconstructed 'at the touch of a button' after updating the model. They got part of the system working - the back end was the hard part. They used object property relationship modelling to hold the high level requirements. However, what ISDOS did was far in advance of the Object Management Group (who use the word 'object' in a very different way). OMG's (and with it UML's) fundamental weakness is that the work is not founded sufficiently well in linguistic terms to be able to perform detailed and automatic completeness and consistency checks on the high level and subsequent models. I wish Brunel the best of luck in handling that fundamental flaw in OMG and UML.