Software Process Measurement

Anne attended this talk on Software Metrics at Lero today.

The speaker was Jari Soini from Tampere University of Technology in Finland. He is in the process of writing up his PhD and because of his interest in Software Process, has connections with Ita Richardson, hence his visit to Lero.

Jari is part of a research group whose aim is to provide a metrics database to assist SMEs in Finland define measurement objectives, i.e. what can and should be measured in the software development process. Knowledge sharing is a key motivator in designing the database. The SMEs are finding measurement difficult and want to learn from each other. Metrics have been gathered from over 40 companies through a combination of interviews and a detailed questionnaire. The metrics database has a web based interface which is designed by the research team with continuous feedback and evaluation from the user group.

There was interesting discussion around potential bias in the evaluation of metrics to be included in the database. The research group doesn't identify who decides within each SME what should be measured; who completes the questionaire; who judges questionaire categories such as reliability &accuracy; who are the users of the system.

As part of the data gathering process 3 categories of beneficiary are identified: upper management; project management; & software engineers. One of the outcomes of the research was that the primary beneficiaries of the metrics are upper management (82%), with only 7% aimed at software engineers as key beneficiaries, despite the fact that software engineers are tasked with collecting most of the measurements. Work has not been done in this project on the actual or perceived impact on software engineers.

Abstracts relating to the topics that were discussed can be found on the Lero website.