Approximately 1,300 people attended the 7th Annual Software Engineering Process Group (SEPG) Conference on 22 through 25 May 1995 in Boston, Massachusetts. Co-sponsored by the Boston Software Process Improvement Network (SPIN) and the Software Engineering Institute (SEI), this conference, as noted by Julia Allen of the SEI, has grown in attendance from 46 people in 1988 to 1,300 today. Joseph Farinello, of the Boston SPIN and co-chair of the conference, stated that the SEPG conference has become the largest SEI-sponsored event. A major objective of the 1995 conference was to present "hard" data and results of SPI initiatives.

John D. Warner, President of Boeing Computer Services, in one of two keynote speeches, discussed "Software Development and Customer Satisfaction." Mr. Warner said the software industry needs to move toward the quality of other industry products, whereby software organizations will provide guarantees to their customers that their products will be defect-free for 25 years or they fix problems for free. He pointed out that the software industry is slow to mature. In 1988, 88% of all software organizations were at SEI Level 1 (the initial or lowest SEI level). In 1993, 85% of all software organizations were still at Level 1. Mr. Warner described the success of the Boeing 777 aircraft as "three million parts flying in close formation." The 777 contains four million source lines of code, the software systems are all functional, and the 777 is service ready and on schedule. The 777 was designed on CAD/CAM systems and involved thousands of engineers and scientists working in the U.S. and Japan.

Eileen S. Quann, President of Fastrak Training, talked about "Transforming Your Workplace Into a Learning Organization." Her major premise was that the most successful software organizations are those organizations that institutionalize learning. Technology shelf life is now five years, that change is constant, to remain competitive we must engage in professional development activities. Professional development is the process of continuing education for all employees. The learning process must be institutionalized; organizations and individuals must commit to continuous learning and training; individuals, organizations and customers should share the cost of learning; and acquisition strategies must change to provide incentives to companies that train.

At the general session on the second day, a lively debate occurred between panelists Bill Curtis, VP and Chief Scientist of Teraquest Metrics, and James Bach, Senior Staff QA Engineer for Borland, around "The CMM Controversy: Is it a Step Forward or a Step Backward?" This was a continuation of a debate that first began in the September 1994 issue of American Programmer magazine. Mr. Curtis, who suppoprted the CMM, stated the CMM is Total Quality Management-based, which results in repeatability and discipline from organizations that employ CMM, and the empirical data supports the CMM approach. The benefits of the CMM include: professional standards in software engineering; improved schedules, costs, and product quality; agility in fast paced markets; and improved skills from software professionals. Mr. Bach contends that the CMM is a step backward, and criticizes the CMM because he believes there is no formal theoretical basis to the CMM, vague empirical data supports the CMM, the CMM ignores people, and institutionalizing the CMM reduces software professionals' creativity. He stated that the CMM implies that Level 1 organizations are "barbaric" when in fact good work occurs in Level 1 companies, dissenters are assumed to be "afraid of change," higher levels are assumed to be better, and that KPAs (Key Process Areas) are always applicable regardless of industry segment, people or any other factor.

Will Hayes and Dave Zubrow of the SEI, in a presentation titled "Moving On Up: Data and experience Doing CMM-Based SPI", presented some preliminary results on new data about organizations that have been reassessed versus first time assessments. Of the organizations having a first time assessment, commercial organizations (39%) outnumber DoD Contractors (36%). 73% of all organizations having their initial assessment are at Level 1; 25% of those being reassessed remain at Level 1 . Figure 1 graphically depicts the process maturity path for 47 organizations that have been reassessed at least once. For example, of the 14 organizations that started at Level 2 in the first assessment, 13 of them moved to Level 3 in the second assessment. The presenters stated that the average time for organizations to move up one maturity level was 18-30 months.

Dennis Goldenson of the SEI presented the results of the first systematic survey of the benefits of higher maturity levels in his presentation entitled "What Happens After the Appraisal?" Goldenson stated that higher maturity levels resulted in improved product quality, staff morale, customer satisfaction, ability to meet schedule, staff productivity, and the ability to meet budget commitments.

Other presentations at the conference included:

Bill Curtis in his tutorial on "Building a Cost-Benefit Case for Software Process Improvement" stated that many companies are achieving cost benefits of 6 to 1 and higher with process improvement, a 200-300 percent increase in productivity, and a 100 percent decrease in delivered defects. However, two-thirds of all process improvement programs fail. The cause of failure was due to executives not being behind the program, resistance from middle management, and ineffective SEPG leadership.

Donna Dunaway of the SEI provided information concerning the new CMM Based Appraisal (CBA) Internal Process Improvement (IPI) method. The CMM Appraisal Framework (CAF) Maturity Questionnaire is used as a starting point for the IPI assessments. Free form interviews are used along with the questionnaire to obtain improvement information. The new CBA IPI focuses on the goals of the CMM, rating them consistent with the CMM Appraisal Framework. From an IPI standpoint, the activities of the CMM are not prescriptive, but are indicative of fulfillment of the goals of the CMM. Any KPAs that are omitted must be justified in writing, but the major emphasis is goal satisfaction. Ms. Dunaway noted that IPI assessments can be tailored to specific projects or focus only on specified KPAs. Twenty-seven CBA IPIs have been conducted since June 1994 and there are currently 116 Lead Assessors in 51 organizations. The SEI continues to gather data, review and examine potential revisions for the CBA IPI method.

Ray Madachy of Litton Data Systems, in discussing the benefits of inspections, reported that for every person-hour of effort applied to inspections, 2.3 person-hours are saved in testing, and inspections reduced by 2/3 the number of defects detected during integration.

Version 2 of the CMM is planned for the 1996-1998 time frame.

Use these buttons to navigate between newsletter articles.
[Previous Article] [Next Article] [In This Issue]