• E-mail
• Print
• Comment
• Font Size
• Digg
• del.icio.us
• Discuss article

Software Process Modeling

Posted on: Thursday, 5 January 2006, 06:00 CST

By Ferguson, Robert

Software Process Modeling Silvia T. Acua and Natalia Juristo. 2005. Springer (http V/www.springer-ny .com). 208 pages. ISBN 0-387- 24261-9. IGSQE Body of Knowledge area: Software Engineering Processes

Software Process Modeling is a collection of papers intending to address the steady pace of change in how to develop software, how to organize to do the work, and how to model the various aspects of the work. It is always a challenge to read and review this type of book. The eight papers have a different focus, different writing style, and will appeal (or not) to different audiences. I chose to review those papers that I thought were the most interesting, although that means that several get very short shrift indeed.

Two of the papers are about the open-source movement. Walt Scacchi analyzes the movement by examining the social network, and Jacques Longchamp describes the application of the Software Process Engineering Metamodel (SPEM) to open-source development. There are certain distinctive community attributes to the open-source movement. There are stages of inclusion from lurker-to-observer-to- contributor-to-elder. There is more or less formality to the progression depending on several factors including size of the community, social value of progression, utility of the end product, and others.

The SPEM model uses UML as its primary notation. The model did help to build an abstraction of the processes and helped to narrow the primary concerns. Use of the model was primarily focused on the release process. SPEM may be a useful communications tool, but I expect more from a model. I should be able to make some decisions based on the model by identifying points of stress and exercising them. Does the model break or suggest a failure when the actual system fails? When I add an element to SPEM, are other connections also obvious?

Two of the papers used numerical simulation tools to understand the software process. Ray Madachy and Barry Boehm analyzed the tradeoffs between dependability and defect prevention, and also analyzed the tradeoffs between market losses and development costs. They suggest that modeling such relationships is difficult because agreement on the definition of dependability is difficult so there are little data. At the same time they suggest that such simulations can be used to justify the addition of needed quality steps in product development.

Lehman, Kahen, and Ramil also used a numerical simulation for managing software evolution. Their results suggest that each unit of progressive work will require some amount of antiregressive work. One could interpret this as saying the bigger the system gets, the more maintenance work is required. Of course, the challenge is calibrating the work so one can determine exactly how much is required. The inclusion of some sensitivity analysis was one of the best parts of the paper. Every paper on simulation needs to discuss the sensitivity of the model to changes in assumptions and input data.

For real researchers, such books serve as an introduction to the larger literature of the subject. Each paper covers an element of the process modeling arena. Each also includes a large bibliography of references for pursuing the researchers and subject further.

Reviewed by Robert Ferguson

rwf@sei.cmu.edu

Copyright American Society for Quality Dec 2005

Source: Software Quality Professional

More News in this Category