DSpace

Nowadays, software is getting more complex due to upcoming requirements from customers. It is therefore necessary to handle the variants of a software system’s different components and to know what customer uses which component and/or variant. In the area of public administrations, the requirements differ as they are not just influenced by the needs of a company but also by laws and regulations. The purpose of variant management is to handle customer-specific parts for different customers within one system. In order to handle such variants, different concepts can be applied. However, although variant management can be used for every kind of software, it needs to be adapted to the specific case of software for public administrations. This research thesis illustrates how feature trees can be extended and adapted to represent the elements of the specific case of variant management for public administration software. The starting point are software product lines (SPL) with the concept of feature trees for describing a system. Adding elements like influencers or rules to the concept of feature trees and extending it with new relations such as requires shows that it is possible to model variants in an area with specific requirements. Modelling a system with the extended feature model allows finding and retrieving variants for individual customers. A model based on feature trees for an application used in public administrations with all relevant features and relations is developed. The investigated application is a message based register with data about residents on cantonal level. The development is based on results of interviews with employees of a company that develops software for public administrations. In order to evaluate the model, actual results of queries are compared to expected results. Queries are used to retrieve information from the modelled application based on a specific syntax. To find out if the approach could be applicable and useful in practice, some of the practitioners are consulted.

Master of Science

Listen

Filter

Einstellungen

Sortieren nach

Ergebnisse pro Seite

Ergebnisse nach Hochschule und Institut