Prof. Dr. F. Schweiggert, University of Ulm
General:
When I left the academic field in 1980 I worked in the department for Software Quality Assurance for 4 years in a company which developed military systems (MIL-Standards, AQAP-13).
After I became a professor for Computer Science at University of Ulm in 1984 I started a course in Software Quality Assurance (later in Software Quality Management - of course) for "softworkers" in industry, which I do twice a year until today.
From 1986 until 1989 I had a project with a liability insurance company; they wanted to put an insurance policy for financial damages caused by defects in information processing. We developed a method to evaluate the potential risks.
I had a lot of smaller projects with industry in the field of Software Quality, e.g. with IBM in defect causal analysis.
In 1993 and 1994 I had a cooperation with a division (about 500 software engineers and about 300 electronic engineers) of a great industrial company in Germany. One main objective was to reduce life cycle time or as they said to increase productivity. We began to analyze the processes:
One approach was to study the project dynamics (see the book of Gerald Weinberg or Abdel-Hamid) which of course was very intuitive but finally successful. Based on a lot of interviews we tried in another step to estimate the distribution of efforts the engineers (only a small group) spent for several task (e.g. communication, writing useful documentation / "useless" documentation, reading manuals, and so on. Finally we had a lot of information to do some steps towards improvement with respect to life cycle time reduction.
One approach was to introduce Fagan's inspection method - but how? I did about 15 workshops (each 3 days and each with 20 engineers). The method was trained, but the main task was to give the right understanding for why to do this, for why to collect failure and effort data.
Conclusion:
I like to join the workshop to learn more about the utilization of data from people in industry for I believe that this is really necessary for process improvement. But perhaps especially here in Germany it is very difficult to convince managers as well as engineers in small companies that this is a sound way out problems they have (but perhaps not recognize). So what is a successful way for introducing quantitative methods? What are the experiences of others? I just can add what I did see, perhaps with an academic view but without any commercial interests.