empirical studies as a vehicle for software engineering projects claes wohlin dept. of communication...
TRANSCRIPT
Empirical Studies as a Vehicle for Software Engineering Projects
Claes Wohlin
Dept. of Communication Systems
Lund University, Sweden
http://www.telecom.lth.se/SERG/
2
Software Engineering course
Content 1(2) Lectures based on part of Software Engineering book. Four questions on written exam. Group study (5 students): Requirements and Specifications, Verification and Validation or Management and Evolution. The work is presented in a seminar and reviewed by another group. In addition, one question for each area on the written exam. Two of the questions should be answered.
3
Software Engineering course
Content 2(2) Project work in group (same group with 5 students)
Characteristics: Fictitious company with a problem The project is supposed to be problem-driven Based on data generated from an empirical study and
industrial data Limited number of hours
4
Information for the project
General project description Organisational description for the fictitious company Two different project descriptions
Lead-time and subjective data User-oriented scenarios as a means for reviews
5
Start-up of projects
All students participated in two empirical studies. The data is provided to the students. The students:
are assumed to be consultants. are given a problem statement from the customer.
meet the customer three times: kick-off, intermediate meeting and executive briefing.
have limited access to an expert. write a project plan.
6
The Empirical Study:Lead-Time and Subjective Data
The students are asked to judge, using a pairwise comparison technique, which subjective factor is more important for a short lead-time. The students are assumed to be “experts”.
This data together with industrial data from 12 software projects is input to the project. The industrial data consists of scores (five-point scale) for 10 subjective factors and lead times.
7
Expected output
Given the data from the empirical study and industry, the students are expected to deliver the following:
Analysis of actual outcome from 12 projects (data set 1) Analysis of knowledge of the “experts” (data set 2) Comparison and analysis of the difference and
similarities between data sets 1 and 2. Method proposal for using subjective factors to judge
lead-time in projects. Recommendations in this area.
8
Summary
The students: get to work with a challenging software engineering problem. participate in an empirical study. get access to real industrial data. learn to use empirical data.