Regarding the SEMAT article in the May 15 issue: It’s been my experience that software engineering and related projects fall down due to poor planning more often than due to technical factors. Some questions:
1) Is the scope well defined and agreed to by the users and developers?
2) Is the scope defined well enough to permit a realistic estimate of cost and schedule?
3) Do test plans provide a clear and precise set of objectives needed for client signoff?
4) In engineering, costs are typically developed in three stages: budgetary based on similar projects; conceptual based on preliminary engineering; and final based on material and labor take-offs from drawings. What is typical for most software engineering projects?
While it would be good to bring academics into the room, I suggest a greater deficiency is in the lack of user input before, during and after the project. The room needs to be big enough for them.
David M. Weiss
Director, Strategic Business Planning, IEEE