Ph D seminar for the software engineering group

Ph. D. seminar for the software engineering group SEVO - project Anita Gupta anitaash@idi. ntnu. no Norwegian University of Science and Technology 23 rd November 2004

Agenda n Presentation of the projects: SJEF, SITT and EIAO n Goals and metrics n Research methods n Research questions (temporary) n Related work n Future work n Discussion

SJEF n Statoil have called their project for “SJEF”, and SJEF is a new reference architecture which is meant to become the basis for all new software systems. n Issues: ¨ to evaluate the cost efficiency in development, integration and deployment of new systems using SJEF (vision vs reality). ¨ to evaluate the possibilities for reuse (buy vs develop) with SJEF.

SITT n Teleca have called their project for “SITT” and the focus in this project is on simplistic solutions. n Issues are: ¨ how to link Bugzilla (used for bug tracking) to CVS (used for configuration management). ¨ data amount/classifications vs. scalability and simplicity.

EIAO n Høyskolen i Agder have called their project for “EIAO” (European Internet Accessibility Observatory), and the project is about web evolution. ¨ Automatic assessment of website accessibility (ROBACC tool). n Issues ¨ Goals and metrics both for ROBACC tool and web evolution.

Goals and metrics n Come up with goals and software metrics (GQM approach), which will be helpful for the projects. n Come up with goals and software metrics (GQM approach) that will be helpful for us to collect some data on evolution and maybe see some similarities/differences between these projects.

Research methods n For these 3 projects: Case study, we are going to collect data from the project. n For other possible projects: Experiment and data mining.

Reserach questions (temporary) n n What impact does software evolution has on quality attributes and on software processes? Are there any similarities/differences between the projects? What improvements issues are there, and what does it takes to handle these improvements? How can software evolution improve some of the software processes?

Related work n A student has made a summery over the topic of software evolution. He has done a study of changes (evolution) of different software architectures of 2 -3 big industrial systems. The emphasize is on identifying the most important factors which can explain and maybe predict such evolution.

Future work n Establish relations to other companies. n Next semester maybe perform an experiment in the subject ”Software architecture”. n Maybe combine software evolution with the systems the hospitals are using to perform their work. n Maybe combine software evolution with expansion of financial or other systems in Norway.

Discussion n Which goals and software metrics could be useful for collecting data on software evolution? n Are there other research methods that I should be aware of, especially with respect to collecting data on software evolution? n Should I have any other research questions with respect to software evolution and possible software process improvement?