login

Towards a More Precise Definition of Feature Models

Published 1 January 2003
Matthias Riebisch
Citations60

TL;DR

This position paper summarizes the state of the discussion and proposes issues for future development of feature models, a well accepted means for expressing requirements in a domain on an abstract level.

Abstract

Feature models are a well accepted means for expressing requirements in a domain on an abstract level. They are applied to describe variable and common properties of products in a product line, and to derive and validate configurations of software systems. Their industrial importance is increasing rapidly. However, methodical usage and tool support demands for a more precise definition of features, their properties and their relations within a feature model. This position paper summarizes the state of the discussion and proposes issues for future development. Categories of features and types of their attributes and relations are presented. The represented information is limited to a customer point of view onto the feature models without excluding technically detailed features. Connections of features to other models i.e. design, and to implementation elements are given by traceability links. Approaches for graphical representations and data models for feature models are shown. Proposals of attaching additional information for related tasks like product line evolution, scoping, effort estimation, definition of product configurations and documenting are discussed.

Keywords

Computer Science